{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# AIS Workbook\n",
    "### This workbook utilizes AIS Data to establish travel time distributions for iron ore carriers through specified segments of the Great Lakes waterway.\n",
    "#### Source of data https://marinecadastre.gov/ais/\n",
    "#### Index of Universal Transverse Mercator Zone at https://marinecadastre.gov/AIS/AIS%20Documents/UTMZoneMap2014.png\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The U.S. Coast Guard collects Automatic Identification System (AIS) data through an onboard navigation safety device.  The AIS transmits the location and characteristics of vessels in real time. The Bureau of Ocean Energy Management (BOEM) and the National Oceanic and Atmospheric Administration (NOAA) jointly make available some of the historical data from the network of AIS receivers. Information such as location, time, ship type, speed, length, beam, and draft have been extracted from the raw data and prepared for researchers.\n",
    "\n",
    "Data for this analysis were assembled for UTM zones 15-18 covering the period March 2013 to Jan 2018.  Monthly data files were trimmed to only include the Great Lakes Region and appended into a single monthly file.  Monthly files were later appended into a single database for Great Lakes traffic (to be published under Data in Brief).  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# importing required packages\n",
    "\n",
    "import matplotlib.pyplot as plt #for plotting \n",
    "\n",
    "import os\n",
    "import conda\n",
    "\n",
    "conda_file_dir = conda.__file__\n",
    "conda_dir = conda_file_dir.split('lib')[0]\n",
    "proj_lib = os.path.join(os.path.join(conda_dir, 'share'), 'proj')\n",
    "os.environ[\"PROJ_LIB\"] = proj_lib\n",
    "\n",
    "from mpl_toolkits.basemap import Basemap\n",
    "import matplotlib.colors as colors\n",
    "import matplotlib.cm as cmx #for colormap \n",
    "from matplotlib.collections import LineCollection\n",
    "#from matplotlib.patches import Polygon\n",
    "import scipy.stats as stats\n",
    "import itertools\n",
    "\n",
    "import pandas as pd             #for analyzing data\n",
    "import numpy as np              #for numpy mean & accuracy\n",
    "import datetime as dt           #for datetime objects\n",
    "from datetime import datetime, timedelta\n",
    "import seaborn as sns\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "NY2015 loaded with 13228824 lines.\n",
      "NY2016 loaded with 18782779 lines.\n",
      "NY2017 loaded with 16816603 lines.\n",
      "Files appended\n",
      "There are 48828206 original lines of data in the Dataframe.\n",
      "There are 2205 unique vessels in the dataset.\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MMSI</th>\n",
       "      <th>LAT</th>\n",
       "      <th>LON</th>\n",
       "      <th>SOG</th>\n",
       "      <th>VesselName</th>\n",
       "      <th>Status</th>\n",
       "      <th>Draft</th>\n",
       "      <th>Cargo</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BaseDateTime</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2015-03-01 00:00:02</th>\n",
       "      <td>316003124</td>\n",
       "      <td>42.02662</td>\n",
       "      <td>-82.60184</td>\n",
       "      <td>0.0</td>\n",
       "      <td>PELEE ISLANDER</td>\n",
       "      <td>undefined</td>\n",
       "      <td>2.33</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-01 00:00:03</th>\n",
       "      <td>316002238</td>\n",
       "      <td>46.50399</td>\n",
       "      <td>-84.32688</td>\n",
       "      <td>0.0</td>\n",
       "      <td>ANGLIAN LADY</td>\n",
       "      <td>undefined</td>\n",
       "      <td>4.27</td>\n",
       "      <td>52.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-01 00:00:03</th>\n",
       "      <td>316002238</td>\n",
       "      <td>46.50399</td>\n",
       "      <td>-84.32688</td>\n",
       "      <td>0.0</td>\n",
       "      <td>ANGLIAN LADY</td>\n",
       "      <td>undefined</td>\n",
       "      <td>4.27</td>\n",
       "      <td>52.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-01 00:00:03</th>\n",
       "      <td>367063240</td>\n",
       "      <td>46.48492</td>\n",
       "      <td>-84.29752</td>\n",
       "      <td>0.0</td>\n",
       "      <td>SUGAR ISLANDER II</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>0.00</td>\n",
       "      <td>60.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-01 00:00:03</th>\n",
       "      <td>367063240</td>\n",
       "      <td>46.48492</td>\n",
       "      <td>-84.29752</td>\n",
       "      <td>0.0</td>\n",
       "      <td>SUGAR ISLANDER II</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>0.00</td>\n",
       "      <td>60.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                          MMSI       LAT       LON  SOG         VesselName  \\\n",
       "BaseDateTime                                                                 \n",
       "2015-03-01 00:00:02  316003124  42.02662 -82.60184  0.0     PELEE ISLANDER   \n",
       "2015-03-01 00:00:03  316002238  46.50399 -84.32688  0.0       ANGLIAN LADY   \n",
       "2015-03-01 00:00:03  316002238  46.50399 -84.32688  0.0       ANGLIAN LADY   \n",
       "2015-03-01 00:00:03  367063240  46.48492 -84.29752  0.0  SUGAR ISLANDER II   \n",
       "2015-03-01 00:00:03  367063240  46.48492 -84.29752  0.0  SUGAR ISLANDER II   \n",
       "\n",
       "                                     Status  Draft  Cargo  \n",
       "BaseDateTime                                               \n",
       "2015-03-01 00:00:02               undefined   2.33    NaN  \n",
       "2015-03-01 00:00:03               undefined   4.27   52.0  \n",
       "2015-03-01 00:00:03               undefined   4.27   52.0  \n",
       "2015-03-01 00:00:03  under way using engine   0.00   60.0  \n",
       "2015-03-01 00:00:03  under way using engine   0.00   60.0  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Importing Previously processed AIS Data for the Great Lakes \n",
    "#NY2014 = pd.read_csv('./Data/Trimmed_NY2014.csv')\n",
    "#print(f'NY2014 loaded with {len(NY2014)} lines.')\n",
    "NY2015 = pd.read_csv('./Data/Trimmed_NY2015_new.csv')\n",
    "print(f'NY2015 loaded with {len(NY2015)} lines.')\n",
    "NY2016 = pd.read_csv('./Data/Trimmed_NY2016_new.csv')\n",
    "print(f'NY2016 loaded with {len(NY2016)} lines.')\n",
    "NY2017 = pd.read_csv('./Data/Trimmed_NY2017_new.csv')\n",
    "print(f'NY2017 loaded with {len(NY2017)} lines.')\n",
    "AIS = NY2015.append([NY2016, NY2017])\n",
    "print('Files appended')\n",
    "\n",
    "print(f\"There are {len(AIS)} original lines of data in the Dataframe.\")\n",
    "\n",
    "#Renaming select entries to match records in the LPMS\n",
    "AIS['VesselName'] = AIS['VesselName'].str.replace('JOETHOMPSON','JOSEPH H THOMPSON JR')\n",
    "AIS['VesselName'] = AIS['VesselName'].str.replace('CLYDE S VAN ENKEVORT','TUG CLYDE S VANENKEVORT')\n",
    "AIS = AIS.set_index(pd.to_datetime(AIS['BaseDateTime']), drop=True) #set index to Datetime\n",
    "\n",
    "#print(f\"There are {len(AIS)} lines of data with unnamed vessels removed.\")\n",
    "print(f\"There are {len(np.unique(AIS['MMSI']))} unique vessels in the dataset.\")\n",
    "\n",
    "#Reducing the dataframe to only include columns listed below.  \n",
    "AIS=AIS[['MMSI','LAT','LON','SOG','VesselName','Status','Draft','Cargo']]      \n",
    "AIS.head()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Lock Data was obtained from the Soo Locks facility in Sault Ste Marie, MI in its raw form.  Vessel Names and identification numbers have been obscured to protect proprietary information for shipping companies.  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "55342\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>VesselName</th>\n",
       "      <th>Country</th>\n",
       "      <th>Origin</th>\n",
       "      <th>Destination</th>\n",
       "      <th>Code</th>\n",
       "      <th>Tons</th>\n",
       "      <th>Cargo</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BaseDateTime</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2005-03-25 00:01:00</th>\n",
       "      <td>INDIANA HARBOR</td>\n",
       "      <td>840</td>\n",
       "      <td>STURGEON BAY, WI</td>\n",
       "      <td>TWO HARBORS, MN</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>Empty Barges</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2005-03-25 02:56:00</th>\n",
       "      <td>AMERICAN SPIRIT</td>\n",
       "      <td>840</td>\n",
       "      <td>SUPERIOR, WI</td>\n",
       "      <td>GARY, IN</td>\n",
       "      <td>4410</td>\n",
       "      <td>53952</td>\n",
       "      <td>Iron Ore</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2005-03-25 11:25:00</th>\n",
       "      <td>PRESQUE ISLE</td>\n",
       "      <td>840</td>\n",
       "      <td>TWO HARBORS, MN</td>\n",
       "      <td>GARY, IN</td>\n",
       "      <td>4410</td>\n",
       "      <td>55608</td>\n",
       "      <td>Iron Ore</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2005-03-25 13:14:00</th>\n",
       "      <td>CHARLES M BEEGHLY</td>\n",
       "      <td>840</td>\n",
       "      <td>CALUMET HARBOR, IL</td>\n",
       "      <td>PRESQUE ISLE, MI</td>\n",
       "      <td>1100</td>\n",
       "      <td>22287</td>\n",
       "      <td>Coal Lignite</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2005-03-25 14:33:00</th>\n",
       "      <td>LEE A TREGURTHA</td>\n",
       "      <td>840</td>\n",
       "      <td>PORT LAMBTON CAN</td>\n",
       "      <td>DULUTH, MN</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>Empty Barges</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                            VesselName Country              Origin  \\\n",
       "BaseDateTime                                                         \n",
       "2005-03-25 00:01:00     INDIANA HARBOR     840    STURGEON BAY, WI   \n",
       "2005-03-25 02:56:00    AMERICAN SPIRIT     840        SUPERIOR, WI   \n",
       "2005-03-25 11:25:00       PRESQUE ISLE     840     TWO HARBORS, MN   \n",
       "2005-03-25 13:14:00  CHARLES M BEEGHLY     840  CALUMET HARBOR, IL   \n",
       "2005-03-25 14:33:00    LEE A TREGURTHA     840    PORT LAMBTON CAN   \n",
       "\n",
       "                          Destination  Code   Tons         Cargo  \n",
       "BaseDateTime                                                      \n",
       "2005-03-25 00:01:00   TWO HARBORS, MN     1      0  Empty Barges  \n",
       "2005-03-25 02:56:00          GARY, IN  4410  53952      Iron Ore  \n",
       "2005-03-25 11:25:00          GARY, IN  4410  55608      Iron Ore  \n",
       "2005-03-25 13:14:00  PRESQUE ISLE, MI  1100  22287  Coal Lignite  \n",
       "2005-03-25 14:33:00        DULUTH, MN     1      0  Empty Barges  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Importing Maritime Traffic Data from the Soo Locks and removing unused columns \n",
    "LPMS = pd.read_csv('./Data/Combined Soo Locks Data.csv')\n",
    "#Delete unused columns from the original data\n",
    "del LPMS['VESSEL_FUNCTION_TYPE']\n",
    "del LPMS['BARGE_ID']\n",
    "del LPMS['CHMBR_NO']\n",
    "del LPMS['LOCK_RESTRICTION']\n",
    "del LPMS['EROC']\n",
    "del LPMS['RIVER_CODE']\n",
    "del LPMS['VESSEL_NO']\n",
    "#Renaming columns to be more intuitive\n",
    "LPMS = LPMS.rename(columns={\"VESSEL_NAME\":\"VesselName\",'VESSEL_FLAG':'Country', \n",
    "                              'SOL_DATE':'BaseDateTime',\"ORIGIN\":\"Origin\", \"DESTINATION\":\"Destination\", \n",
    "                              \"COMM_CODE\":\"Code\", 'COMMODITY_TONS':'Tons', \"COMM_DESC\":\"Cargo\"})\n",
    "#Renaming select names to match the AIS dataframe\n",
    "LPMS['VesselName'] = LPMS['VesselName'].str.replace('.','')\n",
    "LPMS['VesselName'] = LPMS['VesselName'].str.replace('G3  MARQUIS','G3 MARQUIS')\n",
    "LPMS['VesselName'] = LPMS['VesselName'].str.replace('ST CLAIR ','ST CLAIR')\n",
    "LPMS['VesselName'] = LPMS['VesselName'].str.replace('ALGOMA BUFFALO','BUFFALO')\n",
    "#LPMS['VesselName'] = LPMS['VesselName'].str.replace('DEFIANCE','TUG DEFIANCE') #\n",
    "LPMS['VesselName'] = LPMS['VesselName'].str.replace('CSL ST LAURENT','CSL ST-LAURENT')\n",
    "\n",
    "#Setting timestamp as index and deleting the redundant column\n",
    "LPMS = LPMS.set_index(pd.to_datetime(LPMS['BaseDateTime']), drop=True)\n",
    "del LPMS['BaseDateTime']\n",
    "print(len(LPMS))\n",
    "LPMS.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "11032\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>VesselName</th>\n",
       "      <th>Country</th>\n",
       "      <th>Origin</th>\n",
       "      <th>Destination</th>\n",
       "      <th>Code</th>\n",
       "      <th>Tons</th>\n",
       "      <th>Cargo</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BaseDateTime</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2015-03-25 20:09:00</th>\n",
       "      <td>EDWIN H GOTT</td>\n",
       "      <td>840</td>\n",
       "      <td>MILWAUKEE, WI</td>\n",
       "      <td>TWO HARBORS, MN</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>Empty Barges</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-26 01:12:00</th>\n",
       "      <td>ROGER BLOUGH</td>\n",
       "      <td>840</td>\n",
       "      <td>ERIE, PA</td>\n",
       "      <td>TWO HARBORS, MN</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>Empty Barges</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-27 18:22:00</th>\n",
       "      <td>JOHN G MUNSON</td>\n",
       "      <td>840</td>\n",
       "      <td>DULUTH, MN</td>\n",
       "      <td>GARY, IN</td>\n",
       "      <td>4410</td>\n",
       "      <td>26837</td>\n",
       "      <td>Iron Ore</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-29 08:05:00</th>\n",
       "      <td>ALGOMA OLYMPIC</td>\n",
       "      <td>124</td>\n",
       "      <td>SARNIA CAN</td>\n",
       "      <td>SUPERIOR, WI</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>Empty Barges</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-30 19:40:00</th>\n",
       "      <td>PHILIP R CLARKE</td>\n",
       "      <td>840</td>\n",
       "      <td>TOLEDO, OH</td>\n",
       "      <td>TWO HARBORS, MN</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>Empty Barges</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                          VesselName Country         Origin      Destination  \\\n",
       "BaseDateTime                                                                   \n",
       "2015-03-25 20:09:00     EDWIN H GOTT     840  MILWAUKEE, WI  TWO HARBORS, MN   \n",
       "2015-03-26 01:12:00     ROGER BLOUGH     840       ERIE, PA  TWO HARBORS, MN   \n",
       "2015-03-27 18:22:00    JOHN G MUNSON     840     DULUTH, MN         GARY, IN   \n",
       "2015-03-29 08:05:00   ALGOMA OLYMPIC     124     SARNIA CAN     SUPERIOR, WI   \n",
       "2015-03-30 19:40:00  PHILIP R CLARKE     840     TOLEDO, OH  TWO HARBORS, MN   \n",
       "\n",
       "                     Code   Tons         Cargo  \n",
       "BaseDateTime                                    \n",
       "2015-03-25 20:09:00     1      0  Empty Barges  \n",
       "2015-03-26 01:12:00     1      0  Empty Barges  \n",
       "2015-03-27 18:22:00  4410  26837      Iron Ore  \n",
       "2015-03-29 08:05:00     1      0  Empty Barges  \n",
       "2015-03-30 19:40:00     1      0  Empty Barges  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#This clips the LPMS data the navigation years of interest.  \n",
    "#A navigation year in the Great Lakes is 25 March to 15 Jan.  I have assigned a 5 days buffer.\n",
    "NY = LPMS.loc[(LPMS.index>'2015-3-20')&(LPMS.index<'2018-1-20')]\n",
    "print(len(NY))\n",
    "NY.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>VesselName</th>\n",
       "      <th>Country</th>\n",
       "      <th>Origin</th>\n",
       "      <th>Destination</th>\n",
       "      <th>Code</th>\n",
       "      <th>Tons</th>\n",
       "      <th>Cargo</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BaseDateTime</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2018-01-13 20:24:00</th>\n",
       "      <td>TUG ANGLIAN LADY</td>\n",
       "      <td>124</td>\n",
       "      <td>SAULT STE MARIE (ABOVE) CAN</td>\n",
       "      <td>DETROIT, MI</td>\n",
       "      <td>5330</td>\n",
       "      <td>10108</td>\n",
       "      <td>Iron and Steel Plates and Sheets</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-01-14 15:03:00</th>\n",
       "      <td>JOSEPH L BLOCK</td>\n",
       "      <td>840</td>\n",
       "      <td>SUPERIOR, WI</td>\n",
       "      <td>BURNS HARBOR, IN</td>\n",
       "      <td>4410</td>\n",
       "      <td>31360</td>\n",
       "      <td>Iron Ore</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-01-14 16:16:00</th>\n",
       "      <td>AMERICAN MARINER</td>\n",
       "      <td>840</td>\n",
       "      <td>TWO HARBORS, MN</td>\n",
       "      <td>DETROIT, MI</td>\n",
       "      <td>4410</td>\n",
       "      <td>30838</td>\n",
       "      <td>Iron Ore</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-01-15 09:49:00</th>\n",
       "      <td>JOYCE L VANENKEVORT</td>\n",
       "      <td>840</td>\n",
       "      <td>TWO HARBORS, MN</td>\n",
       "      <td>CONNEAUT, OH</td>\n",
       "      <td>4410</td>\n",
       "      <td>32206</td>\n",
       "      <td>Iron Ore</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2018-01-15 14:23:00</th>\n",
       "      <td>EDWIN H GOTT</td>\n",
       "      <td>840</td>\n",
       "      <td>DETROIT, MI</td>\n",
       "      <td>DULUTH, MN</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>Empty Barges</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                              VesselName Country                       Origin  \\\n",
       "BaseDateTime                                                                    \n",
       "2018-01-13 20:24:00     TUG ANGLIAN LADY     124  SAULT STE MARIE (ABOVE) CAN   \n",
       "2018-01-14 15:03:00       JOSEPH L BLOCK     840                 SUPERIOR, WI   \n",
       "2018-01-14 16:16:00     AMERICAN MARINER     840              TWO HARBORS, MN   \n",
       "2018-01-15 09:49:00  JOYCE L VANENKEVORT     840              TWO HARBORS, MN   \n",
       "2018-01-15 14:23:00         EDWIN H GOTT     840                  DETROIT, MI   \n",
       "\n",
       "                          Destination  Code   Tons  \\\n",
       "BaseDateTime                                         \n",
       "2018-01-13 20:24:00       DETROIT, MI  5330  10108   \n",
       "2018-01-14 15:03:00  BURNS HARBOR, IN  4410  31360   \n",
       "2018-01-14 16:16:00       DETROIT, MI  4410  30838   \n",
       "2018-01-15 09:49:00      CONNEAUT, OH  4410  32206   \n",
       "2018-01-15 14:23:00        DULUTH, MN     1      0   \n",
       "\n",
       "                                                Cargo  \n",
       "BaseDateTime                                           \n",
       "2018-01-13 20:24:00  Iron and Steel Plates and Sheets  \n",
       "2018-01-14 15:03:00                          Iron Ore  \n",
       "2018-01-14 16:16:00                          Iron Ore  \n",
       "2018-01-15 09:49:00                          Iron Ore  \n",
       "2018-01-15 14:23:00                      Empty Barges  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "NY.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There were 2862 shipments of iron ore in this time period.\n",
      "83 unique vessels carried iron ore.\n"
     ]
    }
   ],
   "source": [
    "iron = NY.loc[NY.Cargo=='Iron Ore'] #subsets LPMS data to only include vessels that have carried iron ore in the specified year\n",
    "print(f'There were {len(iron)} shipments of iron ore in this time period.')\n",
    "vessels = iron.VesselName.unique() #list of iron carrying vessels\n",
    "print(f'{len(vessels)} unique vessels carried iron ore.')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Index(['AMERICAN CENTURY', 'AMERICAN INTEGRITY', 'AMERICAN SPIRIT', 'BUFFALO',\n",
       "       'BURNS HARBOR', 'CASON J CALLAWAY', 'CSL ASSINIBOINE', 'CSL LAURENTIEN',\n",
       "       'CSL NIAGARA', 'EDGAR B SPEER', 'EDWIN H GOTT', 'HERBERT C JACKSON',\n",
       "       'HON JAMES L OBERSTAR', 'JAMES R BARKER', 'JOHN G MUNSON',\n",
       "       'JOSEPH L BLOCK', 'JOYCE L VANENKEVORT', 'KAYE E BARKER',\n",
       "       'LEE A TREGURTHA', 'MESABI MINER', 'PHILIP R CLARKE', 'PRESQUE ISLE',\n",
       "       'ROGER BLOUGH', 'RT HON PAUL J MARTIN', 'SAM LAUD', 'STEWART J CORT',\n",
       "       'THUNDER BAY', 'TUG CLYDE S VANENKEVORT', 'VICTORY',\n",
       "       'WALTER J MCCARTHY JR'],\n",
       "      dtype='object', name='VesselName')"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Creating list of vessels that carry iron ore\n",
    "vessels = iron.VesselName.unique()\n",
    "\n",
    "#Restricting the list of vessels to those that have made 10 or more trips with iron ore.\n",
    "#I wanted the data to mostly reflect vessesl that are making the back and forth runs between lakes.  \n",
    "count=iron.groupby(iron.VesselName)['Tons'].count()\n",
    "shipper = count.loc[(count>30)==True]\n",
    "vessels=shipper.index #list of vessels with >10 iron trips\n",
    "print(len(vessels))\n",
    "vessels\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 2271 trips of iron ore in the dataset.\n",
      "There are 10 vessels larger than 65,000 Tons.\n",
      "There are 10 vessels smaller than 35,000 Tons.\n",
      "There are 10 vessels between 35,000 and 65,000 Tons.\n",
      "There are 837 records for big vessels.\n",
      "There are 606 records for medium vessels.\n",
      "There are 828 records for small vessels.\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>VesselName</th>\n",
       "      <th>Country</th>\n",
       "      <th>Origin</th>\n",
       "      <th>Destination</th>\n",
       "      <th>Code</th>\n",
       "      <th>Tons</th>\n",
       "      <th>Cargo</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BaseDateTime</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2015-04-02 13:20:00</th>\n",
       "      <td>EDWIN H GOTT</td>\n",
       "      <td>840</td>\n",
       "      <td>TWO HARBORS, MN</td>\n",
       "      <td>GARY, IN</td>\n",
       "      <td>4410</td>\n",
       "      <td>66157</td>\n",
       "      <td>Iron Ore</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-04-09 22:28:00</th>\n",
       "      <td>STEWART J CORT</td>\n",
       "      <td>840</td>\n",
       "      <td>SUPERIOR, WI</td>\n",
       "      <td>BURNS HARBOR, IN</td>\n",
       "      <td>4410</td>\n",
       "      <td>62526</td>\n",
       "      <td>Iron Ore</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-04-12 02:28:00</th>\n",
       "      <td>AMERICAN INTEGRITY</td>\n",
       "      <td>840</td>\n",
       "      <td>SILVER BAY, MN</td>\n",
       "      <td>CLEVELAND, OH</td>\n",
       "      <td>4410</td>\n",
       "      <td>63499</td>\n",
       "      <td>Iron Ore</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-04-13 04:46:00</th>\n",
       "      <td>JAMES R BARKER</td>\n",
       "      <td>840</td>\n",
       "      <td>DULUTH, MN</td>\n",
       "      <td>INDIANA HARBOR, IN</td>\n",
       "      <td>4410</td>\n",
       "      <td>66136</td>\n",
       "      <td>Iron Ore</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-04-13 09:05:00</th>\n",
       "      <td>EDGAR B SPEER</td>\n",
       "      <td>840</td>\n",
       "      <td>TWO HARBORS, MN</td>\n",
       "      <td>GARY, IN</td>\n",
       "      <td>4410</td>\n",
       "      <td>67376</td>\n",
       "      <td>Iron Ore</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                             VesselName Country           Origin  \\\n",
       "BaseDateTime                                                       \n",
       "2015-04-02 13:20:00        EDWIN H GOTT     840  TWO HARBORS, MN   \n",
       "2015-04-09 22:28:00      STEWART J CORT     840     SUPERIOR, WI   \n",
       "2015-04-12 02:28:00  AMERICAN INTEGRITY     840   SILVER BAY, MN   \n",
       "2015-04-13 04:46:00      JAMES R BARKER     840       DULUTH, MN   \n",
       "2015-04-13 09:05:00       EDGAR B SPEER     840  TWO HARBORS, MN   \n",
       "\n",
       "                            Destination  Code   Tons     Cargo  \n",
       "BaseDateTime                                                    \n",
       "2015-04-02 13:20:00            GARY, IN  4410  66157  Iron Ore  \n",
       "2015-04-09 22:28:00    BURNS HARBOR, IN  4410  62526  Iron Ore  \n",
       "2015-04-12 02:28:00       CLEVELAND, OH  4410  63499  Iron Ore  \n",
       "2015-04-13 04:46:00  INDIANA HARBOR, IN  4410  66136  Iron Ore  \n",
       "2015-04-13 09:05:00            GARY, IN  4410  67376  Iron Ore  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Trisecting data by vessel class \n",
    "begin= pd.datetime(year=2015, month = 3, day = 25) #beginning of navigation year 2005\n",
    "end  = pd.datetime(year=2018, month = 1, day = 15) # end of navigation year 2017\n",
    "#subsets iron ore data for 44 vessels operating in navigation years 2005-2017\n",
    "iron=iron.loc[((iron.VesselName.isin(vessels))&(iron.index>begin)&(iron.index<end))] \n",
    "\n",
    "print(f'There are {len(iron)} trips of iron ore in the dataset.')\n",
    "MaxLoad = iron.groupby(iron.VesselName)['Tons'].max() #Max Load carried for each vessel in the data\n",
    "\n",
    "Bigs = MaxLoad.loc[(MaxLoad>65000)==True]\n",
    "print(f'There are {len(Bigs)} vessels larger than 65,000 Tons.')\n",
    "Smalls = MaxLoad.loc[(MaxLoad<35000)==True]\n",
    "print(f'There are {len(Smalls)} vessels smaller than 35,000 Tons.')\n",
    "Mediums = MaxLoad.loc[((MaxLoad>35000)&(MaxLoad<65000))==True]\n",
    "print(f'There are {len(Mediums)} vessels between 35,000 and 65,000 Tons.')\n",
    "\n",
    "\n",
    "BIG= iron.loc[iron.VesselName.isin(Bigs.index)]\n",
    "print(f'There are {len(BIG)} records for big vessels.')\n",
    "MEDIUM= iron.loc[iron.VesselName.isin(Mediums.index)]\n",
    "print(f'There are {len(MEDIUM)} records for medium vessels.')\n",
    "SMALL= iron.loc[iron.VesselName.isin(Smalls.index)]\n",
    "print(f'There are {len(SMALL)} records for small vessels.')\n",
    "BIG.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([366971350, 316001635, 366904880, 366904910, 366930730, 366971360,\n",
       "       367082230, 366972030, 366971370, 366983440, 367480260, 367050550,\n",
       "       366904950, 366904930, 367057570, 366971330, 366901880, 366971340,\n",
       "       366905890, 316029000, 367120990, 316023339, 366904890, 316001637,\n",
       "       367121010, 366906610, 366938760, 316001633, 366938740, 338888000])"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Identifying MMSI for each of the iron vessels.  The MMSI is necessary to identify ships in the AIS data before 2015.  \n",
    "#Three vessels have no listed MMSI, and two have multiple listings \n",
    "ais=AIS.loc[AIS.VesselName.isin(vessels)]\n",
    "MMSIs=ais.MMSI.unique()\n",
    "print(len(MMSIs))\n",
    "MMSIs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"Iron=pd.DataFrame()\\nfor MMSI in MMSIs:\\n    subset = AIS.loc[(AIS.MMSI==MMSI)]\\n    #creates timedelta between each point in the dataframe for given vessel, units of hours\\n    subset['Duration']=subset.index.to_series().diff().astype('timedelta64[h]') \\n    subset['from_LAT']=subset.LAT.shift() # this records the last position from which the timedelta is calculated\\n    subset['from_LON']=subset.LON.shift() # this records the last position from which the timedelta is calculated\\n    #filling Vessel Names for nan values in the AIS data\\n    subset.loc[:,'VesselName']=subset.loc[:,'VesselName'].bfill() # backfill\\n    subset.loc[:,'VesselName']=subset.loc[:,'VesselName'].ffill() # forward fill\\n    #subsets the data further for dt > specified # hours.  \\n    subset=subset.loc[(subset.Duration>0.5)]\\n    Iron = Iron.append(subset)\\n    #print(f'{MMSI} has {len(subset)} data points.')\\nIron = Iron.sort_index()\\nprint(len(Iron))\\nIron.head()\\n\\n#resetting vessel list to names.  I prefer to use names in the analysis because it is more intuitive.  MMSI would also work.    \\nvessels=Iron.VesselName.unique()\\nprint(f'There are {len(vessels)} vessels that we will analyze movements.')\\nvessels\\n\""
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#This cell is written to fill VesselName data for AIS record prior to 2015.  \n",
    "#IT DOES NOT WORK - MMSIs were encrypted by the USCG until 2015 and the records do not match those for 2015-2017.\n",
    "#The Expansion of this dataset into those earlier years require the encryption key from the USCG.  \n",
    "\n",
    "'''Iron=pd.DataFrame()\n",
    "for MMSI in MMSIs:\n",
    "    subset = AIS.loc[(AIS.MMSI==MMSI)]\n",
    "    #creates timedelta between each point in the dataframe for given vessel, units of hours\n",
    "    subset['Duration']=subset.index.to_series().diff().astype('timedelta64[h]') \n",
    "    subset['from_LAT']=subset.LAT.shift() # this records the last position from which the timedelta is calculated\n",
    "    subset['from_LON']=subset.LON.shift() # this records the last position from which the timedelta is calculated\n",
    "    #filling Vessel Names for nan values in the AIS data\n",
    "    subset.loc[:,'VesselName']=subset.loc[:,'VesselName'].bfill() # backfill\n",
    "    subset.loc[:,'VesselName']=subset.loc[:,'VesselName'].ffill() # forward fill\n",
    "    #subsets the data further for dt > specified # hours.  \n",
    "    subset=subset.loc[(subset.Duration>0.5)]\n",
    "    Iron = Iron.append(subset)\n",
    "    #print(f'{MMSI} has {len(subset)} data points.')\n",
    "Iron = Iron.sort_index()\n",
    "print(len(Iron))\n",
    "Iron.head()\n",
    "\n",
    "#resetting vessel list to names.  I prefer to use names in the analysis because it is more intuitive.  MMSI would also work.    \n",
    "vessels=Iron.VesselName.unique()\n",
    "print(f'There are {len(vessels)} vessels that we will analyze movements.')\n",
    "vessels\n",
    "'''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>West</th>\n",
       "      <th>East</th>\n",
       "      <th>North</th>\n",
       "      <th>South</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Two Harbors</th>\n",
       "      <td>-91.713</td>\n",
       "      <td>-91.601</td>\n",
       "      <td>47.064</td>\n",
       "      <td>46.958</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Silver Bay</th>\n",
       "      <td>-91.277</td>\n",
       "      <td>-91.199</td>\n",
       "      <td>47.292</td>\n",
       "      <td>47.239</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Duluth-Superior</th>\n",
       "      <td>-92.160</td>\n",
       "      <td>-91.852</td>\n",
       "      <td>46.926</td>\n",
       "      <td>46.633</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Presque Isle</th>\n",
       "      <td>-87.395</td>\n",
       "      <td>-87.357</td>\n",
       "      <td>46.582</td>\n",
       "      <td>46.561</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>St Marys River</th>\n",
       "      <td>-84.996</td>\n",
       "      <td>-83.955</td>\n",
       "      <td>46.770</td>\n",
       "      <td>46.107</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   West    East   North   South\n",
       "Two Harbors     -91.713 -91.601  47.064  46.958\n",
       "Silver Bay      -91.277 -91.199  47.292  47.239\n",
       "Duluth-Superior -92.160 -91.852  46.926  46.633\n",
       "Presque Isle    -87.395 -87.357  46.582  46.561\n",
       "St Marys River  -84.996 -83.955  46.770  46.107"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "features=pd.DataFrame(columns=['West','East','North','South'])\n",
    "features.loc['Two Harbors']    =[-91.713,-91.601,47.064,46.958]\n",
    "features.loc['Silver Bay']     =[-91.277,-91.199,47.292,47.239]\n",
    "features.loc['Duluth-Superior']=[-92.160,-91.852,46.926,46.633]\n",
    "features.loc['Presque Isle']   =[-87.395,-87.357,46.582,46.561]\n",
    "\n",
    "#features.loc['St Marys River'] =[-84.584,-83.955,46.561,46.107]\n",
    "features.loc['St Marys River'] =[-84.996,-83.955,46.770,46.107] #including Whitefish Bay\n",
    "features.loc['St Clair']       =[-82.466,-82.375,43.015,43.009]\n",
    "features.loc['Detroit River']  =[-83.219,-83.062,42.075,42.064]\n",
    "features.loc['W Lake Erie']    =[-82.671,-82.572,42.057,41.406]\n",
    "\n",
    "features.loc['Welland Canal']  =[-79.261,-79.182,43.220,42.868]\n",
    "features.loc['Mackinaw Straits']=[-84.753,-84.725,45.857,45.766]\n",
    "#features.loc['Whitefish Bay']  =[-84.996,-84.541,46.770,46.763]\n",
    "features.loc['S.Lake Michigan']=[-87.548,-86.847,41.750,41.743]\n",
    "\n",
    "features.loc['E.Lake Ontario'] =[-76.599,-76.592,44.274,43.421]\n",
    "features.loc['Thunder Bay']    =[-89.267,-89.092,48.476,48.311]\n",
    "features.loc['Nanticoke']      =[-80.054,-80.029,42.802,42.766]\n",
    "features.loc['Hamilton']       =[-79.780,-79.802,43.262,43.303]\n",
    "\n",
    "features.loc['Indiana Harbor'] =[-87.496,-87.429,41.682,41.641]\n",
    "features.loc['Gary']           =[-87.329,-87.319,41.628,41.609]\n",
    "features.loc['Burns Harbor']   =[-87.153,-87.144,41.647,41.634]\n",
    "features.loc['Detroit']        =[-83.110,-83.106,42.281,42.278]\n",
    "\n",
    "features.loc['Dearborn']       =[-83.161,-83.153,42.307,42.297]\n",
    "features.loc['Toledo']         =[-83.543,-83.333,41.769,41.460]\n",
    "features.loc['Cleveland']      =[-81.725,-81.663,41.514,41.460]\n",
    "features.loc['Ashtabula']      =[-80.804,-80.781,41.919,41.878]\n",
    "features.loc['Conneaut']       =[-80.598,-80.540,42.008,41.951]\n",
    "\n",
    "features.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AMERICAN CENTURY has 351627 data points.\n",
      "AMERICAN INTEGRITY has 353867 data points.\n",
      "AMERICAN SPIRIT has 288205 data points.\n",
      "BUFFALO has 399908 data points.\n",
      "BURNS HARBOR has 431646 data points.\n",
      "CASON J CALLAWAY has 190759 data points.\n",
      "CSL ASSINIBOINE has 188342 data points.\n",
      "CSL LAURENTIEN has 230878 data points.\n",
      "CSL NIAGARA has 218725 data points.\n",
      "EDGAR B SPEER has 327319 data points.\n",
      "EDWIN H GOTT has 310941 data points.\n",
      "HERBERT C JACKSON has 150046 data points.\n",
      "HON JAMES L OBERSTAR has 339177 data points.\n",
      "JAMES R BARKER has 254150 data points.\n",
      "JOHN G MUNSON has 121463 data points.\n",
      "JOSEPH L BLOCK has 237300 data points.\n",
      "JOYCE L VANENKEVORT has 266940 data points.\n",
      "KAYE E BARKER has 266818 data points.\n",
      "LEE A TREGURTHA has 234481 data points.\n",
      "MESABI MINER has 259098 data points.\n",
      "PHILIP R CLARKE has 285600 data points.\n",
      "PRESQUE ISLE has 320117 data points.\n",
      "ROGER BLOUGH has 249487 data points.\n",
      "RT HON PAUL J MARTIN has 206186 data points.\n",
      "SAM LAUD has 252461 data points.\n",
      "STEWART J CORT has 273407 data points.\n",
      "THUNDER BAY has 269086 data points.\n",
      "TUG CLYDE S VANENKEVORT has 79822 data points.\n",
      "VICTORY has 319265 data points.\n",
      "WALTER J MCCARTHY JR has 287467 data points.\n",
      "7964588\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MMSI</th>\n",
       "      <th>LAT</th>\n",
       "      <th>LON</th>\n",
       "      <th>SOG</th>\n",
       "      <th>VesselName</th>\n",
       "      <th>Status</th>\n",
       "      <th>Draft</th>\n",
       "      <th>Cargo</th>\n",
       "      <th>Duration</th>\n",
       "      <th>from_LAT</th>\n",
       "      <th>from_LON</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BaseDateTime</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2015-03-01 03:14:00</th>\n",
       "      <td>366971350</td>\n",
       "      <td>41.69872</td>\n",
       "      <td>-83.45046</td>\n",
       "      <td>0.1</td>\n",
       "      <td>PHILIP R CLARKE</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>8.03</td>\n",
       "      <td>70.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-02 16:50:01</th>\n",
       "      <td>316001635</td>\n",
       "      <td>42.88318</td>\n",
       "      <td>-79.24848</td>\n",
       "      <td>0.0</td>\n",
       "      <td>RT HON PAUL J MARTIN</td>\n",
       "      <td>under way sailing</td>\n",
       "      <td>9.01</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-02 16:51:11</th>\n",
       "      <td>316001635</td>\n",
       "      <td>42.88317</td>\n",
       "      <td>-79.24854</td>\n",
       "      <td>0.0</td>\n",
       "      <td>RT HON PAUL J MARTIN</td>\n",
       "      <td>under way sailing</td>\n",
       "      <td>9.01</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>42.88318</td>\n",
       "      <td>-79.24848</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-02 16:52:13</th>\n",
       "      <td>316001635</td>\n",
       "      <td>42.88316</td>\n",
       "      <td>-79.24854</td>\n",
       "      <td>0.0</td>\n",
       "      <td>RT HON PAUL J MARTIN</td>\n",
       "      <td>under way sailing</td>\n",
       "      <td>9.01</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>42.88317</td>\n",
       "      <td>-79.24854</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-02 16:53:22</th>\n",
       "      <td>316001635</td>\n",
       "      <td>42.88314</td>\n",
       "      <td>-79.24849</td>\n",
       "      <td>0.0</td>\n",
       "      <td>RT HON PAUL J MARTIN</td>\n",
       "      <td>under way sailing</td>\n",
       "      <td>9.01</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>42.88316</td>\n",
       "      <td>-79.24854</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                          MMSI       LAT       LON  SOG            VesselName  \\\n",
       "BaseDateTime                                                                    \n",
       "2015-03-01 03:14:00  366971350  41.69872 -83.45046  0.1       PHILIP R CLARKE   \n",
       "2015-03-02 16:50:01  316001635  42.88318 -79.24848  0.0  RT HON PAUL J MARTIN   \n",
       "2015-03-02 16:51:11  316001635  42.88317 -79.24854  0.0  RT HON PAUL J MARTIN   \n",
       "2015-03-02 16:52:13  316001635  42.88316 -79.24854  0.0  RT HON PAUL J MARTIN   \n",
       "2015-03-02 16:53:22  316001635  42.88314 -79.24849  0.0  RT HON PAUL J MARTIN   \n",
       "\n",
       "                                     Status  Draft  Cargo  Duration  from_LAT  \\\n",
       "BaseDateTime                                                                    \n",
       "2015-03-01 03:14:00  under way using engine   8.03   70.0       NaN       NaN   \n",
       "2015-03-02 16:50:01       under way sailing   9.01    0.0       NaN       NaN   \n",
       "2015-03-02 16:51:11       under way sailing   9.01    0.0       1.0  42.88318   \n",
       "2015-03-02 16:52:13       under way sailing   9.01    0.0       1.0  42.88317   \n",
       "2015-03-02 16:53:22       under way sailing   9.01    0.0       1.0  42.88316   \n",
       "\n",
       "                     from_LON  \n",
       "BaseDateTime                   \n",
       "2015-03-01 03:14:00       NaN  \n",
       "2015-03-02 16:50:01       NaN  \n",
       "2015-03-02 16:51:11 -79.24848  \n",
       "2015-03-02 16:52:13 -79.24854  \n",
       "2015-03-02 16:53:22 -79.24854  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Iron=pd.DataFrame()\n",
    "for vessel in vessels:\n",
    "    subset = ais.loc[ais.VesselName==vessel]\n",
    "    subset['Duration']=subset.index.to_series().diff().astype('timedelta64[m]') \n",
    "    subset['from_LAT']=subset.LAT.shift() # this records the last position from which the timedelta is calculated\n",
    "    subset['from_LON']=subset.LON.shift() # this records the last position from which the timedelta is calculated\n",
    "    print(f'{vessel} has {len(subset)} data points.')\n",
    "    Iron = Iron.append(subset)\n",
    "Iron = Iron.sort_index()\n",
    "print(len(Iron))\n",
    "Iron.head()    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x504 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "data=Iron\n",
    "fig = plt.figure(figsize=(15, 7))\n",
    "m = Basemap(projection='merc',llcrnrlon=-93,llcrnrlat=41,urcrnrlon=-74,urcrnrlat=49.1, lat_ts=0, resolution='l')\n",
    "ax = fig.add_subplot(111)\n",
    "\n",
    "lat = data['LAT'].values\n",
    "lon = data['LON'].values\n",
    "\n",
    "m.scatter(lon, lat, latlon=True, c='r', label='O', s=50)#cmap='Reds', alpha=0.5)\n",
    "m.drawparallels(np.arange(40,60,2),labels=[True,True,False,False],dashes=[2,2])\n",
    "m.drawmeridians(np.arange(-95,-70,5),labels=[False,False,False,True],dashes=[2,2])\n",
    "\n",
    "m.drawcountries(linewidth=1)\n",
    "m.drawcoastlines()\n",
    "m.drawstates()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['Two Harbors', 'Silver Bay', 'Duluth-Superior', 'Presque Isle',\n",
       "       'St Marys River', 'St Clair', 'Detroit River', 'W Lake Erie',\n",
       "       'Welland Canal', 'Mackinaw Straits', 'S.Lake Michigan',\n",
       "       'E.Lake Ontario', 'Thunder Bay', 'Nanticoke', 'Hamilton',\n",
       "       'Indiana Harbor', 'Gary', 'Burns Harbor', 'Detroit', 'Dearborn',\n",
       "       'Toledo', 'Cleveland', 'Ashtabula', 'Conneaut'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "features.index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "155\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Cargo</th>\n",
       "      <th>Draft</th>\n",
       "      <th>Duration</th>\n",
       "      <th>LAT</th>\n",
       "      <th>LON</th>\n",
       "      <th>MMSI</th>\n",
       "      <th>Position</th>\n",
       "      <th>SOG</th>\n",
       "      <th>Status</th>\n",
       "      <th>VesselName</th>\n",
       "      <th>from_LAT</th>\n",
       "      <th>from_LON</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2015-04-15 13:29:14</th>\n",
       "      <td>70.0</td>\n",
       "      <td>10.38</td>\n",
       "      <td>783.0</td>\n",
       "      <td>46.10950</td>\n",
       "      <td>-84.01477</td>\n",
       "      <td>366906610.0</td>\n",
       "      <td>Arrive St Marys River</td>\n",
       "      <td>10.8</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>WALTER J MCCARTHY JR</td>\n",
       "      <td>45.81536</td>\n",
       "      <td>-84.72834</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-04-19 16:50:26</th>\n",
       "      <td>70.0</td>\n",
       "      <td>10.38</td>\n",
       "      <td>2107.0</td>\n",
       "      <td>46.76845</td>\n",
       "      <td>-84.90162</td>\n",
       "      <td>366906610.0</td>\n",
       "      <td>Arrive St Marys River</td>\n",
       "      <td>8.7</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>WALTER J MCCARTHY JR</td>\n",
       "      <td>47.01222</td>\n",
       "      <td>-91.60329</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-04-23 23:08:52</th>\n",
       "      <td>70.0</td>\n",
       "      <td>10.38</td>\n",
       "      <td>342.0</td>\n",
       "      <td>46.10766</td>\n",
       "      <td>-84.01453</td>\n",
       "      <td>366906610.0</td>\n",
       "      <td>Arrive St Marys River</td>\n",
       "      <td>10.6</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>WALTER J MCCARTHY JR</td>\n",
       "      <td>45.81527</td>\n",
       "      <td>-84.72916</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-04-27 22:50:07</th>\n",
       "      <td>70.0</td>\n",
       "      <td>10.38</td>\n",
       "      <td>1507.0</td>\n",
       "      <td>46.76988</td>\n",
       "      <td>-84.90471</td>\n",
       "      <td>366906610.0</td>\n",
       "      <td>Arrive St Marys River</td>\n",
       "      <td>5.1</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>WALTER J MCCARTHY JR</td>\n",
       "      <td>46.77000</td>\n",
       "      <td>-91.85316</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-05-02 10:49:54</th>\n",
       "      <td>70.0</td>\n",
       "      <td>10.38</td>\n",
       "      <td>238.0</td>\n",
       "      <td>46.10943</td>\n",
       "      <td>-84.01445</td>\n",
       "      <td>366906610.0</td>\n",
       "      <td>Arrive St Marys River</td>\n",
       "      <td>11.8</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>WALTER J MCCARTHY JR</td>\n",
       "      <td>45.81613</td>\n",
       "      <td>-84.72865</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-12-11 10:28:45</th>\n",
       "      <td>70.0</td>\n",
       "      <td>10.40</td>\n",
       "      <td>1014.0</td>\n",
       "      <td>46.10752</td>\n",
       "      <td>-84.01467</td>\n",
       "      <td>366906610.0</td>\n",
       "      <td>Arrive St Marys River</td>\n",
       "      <td>11.4</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>WALTER J MCCARTHY JR</td>\n",
       "      <td>43.01278</td>\n",
       "      <td>-82.41584</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-12-14 23:02:27</th>\n",
       "      <td>70.0</td>\n",
       "      <td>10.40</td>\n",
       "      <td>1531.0</td>\n",
       "      <td>46.76843</td>\n",
       "      <td>-84.90160</td>\n",
       "      <td>366906610.0</td>\n",
       "      <td>Arrive St Marys River</td>\n",
       "      <td>11.9</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>WALTER J MCCARTHY JR</td>\n",
       "      <td>46.98976</td>\n",
       "      <td>-91.60309</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-12-19 23:18:36</th>\n",
       "      <td>70.0</td>\n",
       "      <td>10.40</td>\n",
       "      <td>1143.0</td>\n",
       "      <td>46.10826</td>\n",
       "      <td>-84.01460</td>\n",
       "      <td>366906610.0</td>\n",
       "      <td>Arrive St Marys River</td>\n",
       "      <td>9.1</td>\n",
       "      <td>at anchor</td>\n",
       "      <td>WALTER J MCCARTHY JR</td>\n",
       "      <td>43.01399</td>\n",
       "      <td>-82.41614</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-12-23 15:31:17</th>\n",
       "      <td>70.0</td>\n",
       "      <td>10.40</td>\n",
       "      <td>1431.0</td>\n",
       "      <td>46.76912</td>\n",
       "      <td>-84.90166</td>\n",
       "      <td>366906610.0</td>\n",
       "      <td>Arrive St Marys River</td>\n",
       "      <td>12.0</td>\n",
       "      <td>at anchor</td>\n",
       "      <td>WALTER J MCCARTHY JR</td>\n",
       "      <td>47.01153</td>\n",
       "      <td>-91.60700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-12-27 21:35:21</th>\n",
       "      <td>70.0</td>\n",
       "      <td>10.40</td>\n",
       "      <td>254.0</td>\n",
       "      <td>46.10772</td>\n",
       "      <td>-84.01464</td>\n",
       "      <td>366906610.0</td>\n",
       "      <td>Arrive St Marys River</td>\n",
       "      <td>9.1</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>WALTER J MCCARTHY JR</td>\n",
       "      <td>45.81556</td>\n",
       "      <td>-84.72633</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>155 rows × 12 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Cargo  Draft  Duration       LAT       LON         MMSI  \\\n",
       "2015-04-15 13:29:14   70.0  10.38     783.0  46.10950 -84.01477  366906610.0   \n",
       "2015-04-19 16:50:26   70.0  10.38    2107.0  46.76845 -84.90162  366906610.0   \n",
       "2015-04-23 23:08:52   70.0  10.38     342.0  46.10766 -84.01453  366906610.0   \n",
       "2015-04-27 22:50:07   70.0  10.38    1507.0  46.76988 -84.90471  366906610.0   \n",
       "2015-05-02 10:49:54   70.0  10.38     238.0  46.10943 -84.01445  366906610.0   \n",
       "...                    ...    ...       ...       ...       ...          ...   \n",
       "2017-12-11 10:28:45   70.0  10.40    1014.0  46.10752 -84.01467  366906610.0   \n",
       "2017-12-14 23:02:27   70.0  10.40    1531.0  46.76843 -84.90160  366906610.0   \n",
       "2017-12-19 23:18:36   70.0  10.40    1143.0  46.10826 -84.01460  366906610.0   \n",
       "2017-12-23 15:31:17   70.0  10.40    1431.0  46.76912 -84.90166  366906610.0   \n",
       "2017-12-27 21:35:21   70.0  10.40     254.0  46.10772 -84.01464  366906610.0   \n",
       "\n",
       "                                  Position   SOG                  Status  \\\n",
       "2015-04-15 13:29:14  Arrive St Marys River  10.8  under way using engine   \n",
       "2015-04-19 16:50:26  Arrive St Marys River   8.7  under way using engine   \n",
       "2015-04-23 23:08:52  Arrive St Marys River  10.6  under way using engine   \n",
       "2015-04-27 22:50:07  Arrive St Marys River   5.1  under way using engine   \n",
       "2015-05-02 10:49:54  Arrive St Marys River  11.8  under way using engine   \n",
       "...                                    ...   ...                     ...   \n",
       "2017-12-11 10:28:45  Arrive St Marys River  11.4  under way using engine   \n",
       "2017-12-14 23:02:27  Arrive St Marys River  11.9  under way using engine   \n",
       "2017-12-19 23:18:36  Arrive St Marys River   9.1               at anchor   \n",
       "2017-12-23 15:31:17  Arrive St Marys River  12.0               at anchor   \n",
       "2017-12-27 21:35:21  Arrive St Marys River   9.1  under way using engine   \n",
       "\n",
       "                               VesselName  from_LAT  from_LON  \n",
       "2015-04-15 13:29:14  WALTER J MCCARTHY JR  45.81536 -84.72834  \n",
       "2015-04-19 16:50:26  WALTER J MCCARTHY JR  47.01222 -91.60329  \n",
       "2015-04-23 23:08:52  WALTER J MCCARTHY JR  45.81527 -84.72916  \n",
       "2015-04-27 22:50:07  WALTER J MCCARTHY JR  46.77000 -91.85316  \n",
       "2015-05-02 10:49:54  WALTER J MCCARTHY JR  45.81613 -84.72865  \n",
       "...                                   ...       ...       ...  \n",
       "2017-12-11 10:28:45  WALTER J MCCARTHY JR  43.01278 -82.41584  \n",
       "2017-12-14 23:02:27  WALTER J MCCARTHY JR  46.98976 -91.60309  \n",
       "2017-12-19 23:18:36  WALTER J MCCARTHY JR  43.01399 -82.41614  \n",
       "2017-12-23 15:31:17  WALTER J MCCARTHY JR  47.01153 -91.60700  \n",
       "2017-12-27 21:35:21  WALTER J MCCARTHY JR  45.81556 -84.72633  \n",
       "\n",
       "[155 rows x 12 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "for vessel in vessels:\n",
    "    df=Iron.loc[Iron.VesselName==vessel] #subset of data for single vessel\n",
    "    Points=pd.DataFrame()\n",
    "    feature='St Marys River' \n",
    "    #boundaries from the feature matrix\n",
    "    E=features.loc[feature,'East']\n",
    "    W=features.loc[feature,'West']\n",
    "    N=features.loc[feature,'North']\n",
    "    S=features.loc[feature,'South']\n",
    "\n",
    "    df1=df.loc[(df.LAT<=N)&(df.LAT>=S)&(df.LON<=E)&(df.LON>=W)] #subset of data for single geographic feature\n",
    "\n",
    "    #creating a list of unique dates in the index\n",
    "    datelist=np.unique(df1.index.date)\n",
    "    #This eliminates the second of any two contiguos dates.  Necessary when vessels are in a feature overnight\n",
    "    dates=datelist[(datelist-np.roll(datelist,1))!=pd.Timedelta(days=1)] \n",
    "\n",
    "    for date in dates:\n",
    "        #identifying the minimum timestamp for each date and recording it, along with a position description\n",
    "        df2=df1.loc[(df1.index.date==date)]\n",
    "        arrive=df2.loc[min(df2.index)]\n",
    "        arrive['Position']=f'Arrive {feature}'\n",
    "        Points=Points.append(arrive)\n",
    "print(len(Points))\n",
    "Points"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AMERICAN CENTURY has 1158 points.\n",
      "AMERICAN INTEGRITY has 1238 points.\n",
      "AMERICAN SPIRIT has 1106 points.\n",
      "BUFFALO has 995 points.\n",
      "BURNS HARBOR has 1166 points.\n",
      "CASON J CALLAWAY has 1027 points.\n",
      "CSL ASSINIBOINE has 885 points.\n",
      "CSL LAURENTIEN has 921 points.\n",
      "CSL NIAGARA has 918 points.\n",
      "EDGAR B SPEER has 1395 points.\n",
      "EDWIN H GOTT has 1511 points.\n",
      "HERBERT C JACKSON has 886 points.\n",
      "HON JAMES L OBERSTAR has 1388 points.\n",
      "JAMES R BARKER has 1090 points.\n",
      "JOHN G MUNSON has 549 points.\n",
      "JOSEPH L BLOCK has 1032 points.\n",
      "JOYCE L VANENKEVORT has 1269 points.\n",
      "KAYE E BARKER has 1257 points.\n",
      "LEE A TREGURTHA has 1285 points.\n",
      "MESABI MINER has 1197 points.\n",
      "PHILIP R CLARKE has 1007 points.\n",
      "PRESQUE ISLE has 1108 points.\n",
      "ROGER BLOUGH has 1218 points.\n",
      "RT HON PAUL J MARTIN has 856 points.\n",
      "SAM LAUD has 913 points.\n",
      "STEWART J CORT has 1294 points.\n",
      "THUNDER BAY has 985 points.\n",
      "TUG CLYDE S VANENKEVORT has 408 points.\n",
      "VICTORY has 1116 points.\n",
      "WALTER J MCCARTHY JR has 869 points.\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Cargo</th>\n",
       "      <th>Draft</th>\n",
       "      <th>Duration</th>\n",
       "      <th>LAT</th>\n",
       "      <th>LON</th>\n",
       "      <th>MMSI</th>\n",
       "      <th>Position</th>\n",
       "      <th>SOG</th>\n",
       "      <th>Status</th>\n",
       "      <th>VesselName</th>\n",
       "      <th>from_LAT</th>\n",
       "      <th>from_LON</th>\n",
       "      <th>From</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2015-03-01 03:14:00</th>\n",
       "      <td>70.0</td>\n",
       "      <td>8.03</td>\n",
       "      <td>0 days 00:00:00</td>\n",
       "      <td>41.69872</td>\n",
       "      <td>-83.45046</td>\n",
       "      <td>366971350.0</td>\n",
       "      <td>Depart Toledo</td>\n",
       "      <td>0.1</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>PHILIP R CLARKE</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Arrive Toledo</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-01 03:14:00</th>\n",
       "      <td>70.0</td>\n",
       "      <td>8.03</td>\n",
       "      <td>NaT</td>\n",
       "      <td>41.69872</td>\n",
       "      <td>-83.45046</td>\n",
       "      <td>366971350.0</td>\n",
       "      <td>Arrive Toledo</td>\n",
       "      <td>0.1</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>PHILIP R CLARKE</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-02 16:50:01</th>\n",
       "      <td>0.0</td>\n",
       "      <td>9.01</td>\n",
       "      <td>NaT</td>\n",
       "      <td>42.88318</td>\n",
       "      <td>-79.24848</td>\n",
       "      <td>316001635.0</td>\n",
       "      <td>Arrive Welland Canal</td>\n",
       "      <td>0.0</td>\n",
       "      <td>under way sailing</td>\n",
       "      <td>RT HON PAUL J MARTIN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-02 19:26:23</th>\n",
       "      <td>0.0</td>\n",
       "      <td>9.01</td>\n",
       "      <td>0 days 02:36:22</td>\n",
       "      <td>42.88314</td>\n",
       "      <td>-79.24848</td>\n",
       "      <td>316001635.0</td>\n",
       "      <td>Depart Welland Canal</td>\n",
       "      <td>0.0</td>\n",
       "      <td>under way sailing</td>\n",
       "      <td>RT HON PAUL J MARTIN</td>\n",
       "      <td>42.88314</td>\n",
       "      <td>-79.24848</td>\n",
       "      <td>Arrive Welland Canal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-03 11:28:41</th>\n",
       "      <td>70.0</td>\n",
       "      <td>8.03</td>\n",
       "      <td>2 days 08:14:41</td>\n",
       "      <td>41.69873</td>\n",
       "      <td>-83.45055</td>\n",
       "      <td>366971350.0</td>\n",
       "      <td>Arrive Toledo</td>\n",
       "      <td>0.1</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>PHILIP R CLARKE</td>\n",
       "      <td>41.69872</td>\n",
       "      <td>-83.45046</td>\n",
       "      <td>Depart Toledo</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Cargo  Draft        Duration       LAT       LON  \\\n",
       "2015-03-01 03:14:00   70.0   8.03 0 days 00:00:00  41.69872 -83.45046   \n",
       "2015-03-01 03:14:00   70.0   8.03             NaT  41.69872 -83.45046   \n",
       "2015-03-02 16:50:01    0.0   9.01             NaT  42.88318 -79.24848   \n",
       "2015-03-02 19:26:23    0.0   9.01 0 days 02:36:22  42.88314 -79.24848   \n",
       "2015-03-03 11:28:41   70.0   8.03 2 days 08:14:41  41.69873 -83.45055   \n",
       "\n",
       "                            MMSI              Position  SOG  \\\n",
       "2015-03-01 03:14:00  366971350.0         Depart Toledo  0.1   \n",
       "2015-03-01 03:14:00  366971350.0         Arrive Toledo  0.1   \n",
       "2015-03-02 16:50:01  316001635.0  Arrive Welland Canal  0.0   \n",
       "2015-03-02 19:26:23  316001635.0  Depart Welland Canal  0.0   \n",
       "2015-03-03 11:28:41  366971350.0         Arrive Toledo  0.1   \n",
       "\n",
       "                                     Status            VesselName  from_LAT  \\\n",
       "2015-03-01 03:14:00  under way using engine       PHILIP R CLARKE       NaN   \n",
       "2015-03-01 03:14:00  under way using engine       PHILIP R CLARKE       NaN   \n",
       "2015-03-02 16:50:01       under way sailing  RT HON PAUL J MARTIN       NaN   \n",
       "2015-03-02 19:26:23       under way sailing  RT HON PAUL J MARTIN  42.88314   \n",
       "2015-03-03 11:28:41  under way using engine       PHILIP R CLARKE  41.69872   \n",
       "\n",
       "                     from_LON                  From  \n",
       "2015-03-01 03:14:00       NaN         Arrive Toledo  \n",
       "2015-03-01 03:14:00       NaN                   NaN  \n",
       "2015-03-02 16:50:01       NaN                   NaN  \n",
       "2015-03-02 19:26:23 -79.24848  Arrive Welland Canal  \n",
       "2015-03-03 11:28:41 -83.45046         Depart Toledo  "
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#This cell records minimum and maximum (for infrastructure only) times for each ship in each feature\n",
    "\n",
    "Inf = ['Two Harbors','Silver Bay','Duluth-Superior','Presque Isle','St Marys River','Welland Canal',\n",
    "       'Thunder Bay','Nanticoke','Hamilton','Indiana Harbor','Gary','Burns Harbor','Detroit','Dearborn',\n",
    "      'Toledo','Cleveland','Ashtabula','Conneaut']\n",
    "Trips=pd.DataFrame()\n",
    "for vessel in vessels:\n",
    "    df=Iron.loc[Iron.VesselName==vessel] #subset of data for single vessel\n",
    "    Points=pd.DataFrame()\n",
    "    for feature in features.index: \n",
    "        #boundaries from the feature matrix\n",
    "        E=features.loc[feature,'East']\n",
    "        W=features.loc[feature,'West']\n",
    "        N=features.loc[feature,'North']\n",
    "        S=features.loc[feature,'South']\n",
    "\n",
    "        df1=df.loc[(df.LAT<=N)&(df.LAT>=S)&(df.LON<=E)&(df.LON>=W)] #subset of data for single geographic feature\n",
    "        \n",
    "        #creating a list of unique dates in the index\n",
    "        datelist=np.unique(df1.index.date)\n",
    "        #This eliminates the second of any two contiguos dates.  Necessary when vessels are in a feature overnight\n",
    "        dates=datelist[(datelist-np.roll(datelist,1))!=pd.Timedelta(days=1)] \n",
    "       \n",
    "        for date in dates:\n",
    "            #identifying the minimum timestamp for each date and recording it, along with a position description\n",
    "            df2=df1.loc[(df1.index.date==date)]\n",
    "            arrive=df2.loc[min(df2.index)]\n",
    "            arrive['Position']=f'Arrive {feature}'\n",
    "            Points=Points.append(arrive)\n",
    "\n",
    "        #Conditional statement for harbors and lock infrastructure.  Only record a departure time for thos features    \n",
    "        if feature in Inf:\n",
    "            datelist=np.unique(df1.index.date)\n",
    "        #This eliminates the first of any two contiguos dates.  Necessary when vessels are in a feature overnight\n",
    "            dates=datelist[(datelist-np.roll(datelist,-1))!=pd.Timedelta(days=-1)]\n",
    "            for date in dates:\n",
    "            #identifying the minimum timestamp for each date and recording it, along with a position description\n",
    "                df2=df1.loc[(df1.index.date==date)]\n",
    "                depart=df2.loc[max(df2.index)]\n",
    "                depart['Position']=f'Depart {feature}'\n",
    "                Points=Points.append(depart)\n",
    "\n",
    "    print(f'{vessel} has {len(Points)} points.')\n",
    "    Points=Points.sort_index()\n",
    "    Points.Duration=Points.index.to_series().diff() #recalculating trip duration based on the captured data\n",
    "    Points['From']=Points.Position.shift()\n",
    "    Trips=Trips.append(Points)\n",
    "Trips=Trips.sort_index()\n",
    "Trips.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "32047"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(Trips)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4604\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Cargo</th>\n",
       "      <th>Draft</th>\n",
       "      <th>Duration</th>\n",
       "      <th>LAT</th>\n",
       "      <th>LON</th>\n",
       "      <th>MMSI</th>\n",
       "      <th>Position</th>\n",
       "      <th>SOG</th>\n",
       "      <th>Status</th>\n",
       "      <th>VesselName</th>\n",
       "      <th>from_LAT</th>\n",
       "      <th>from_LON</th>\n",
       "      <th>From</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2015-03-26 18:36:48</th>\n",
       "      <td>75.0</td>\n",
       "      <td>9.77</td>\n",
       "      <td>2 days 13:31:44</td>\n",
       "      <td>46.76976</td>\n",
       "      <td>-84.90145</td>\n",
       "      <td>366971370.0</td>\n",
       "      <td>Depart St Marys River</td>\n",
       "      <td>6.8</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>EDWIN H GOTT</td>\n",
       "      <td>46.76789</td>\n",
       "      <td>-84.89990</td>\n",
       "      <td>Arrive St Marys River</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-26 18:46:36</th>\n",
       "      <td>70.0</td>\n",
       "      <td>8.51</td>\n",
       "      <td>2 days 06:05:28</td>\n",
       "      <td>46.76783</td>\n",
       "      <td>-84.89953</td>\n",
       "      <td>366972030.0</td>\n",
       "      <td>Depart St Marys River</td>\n",
       "      <td>9.9</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>ROGER BLOUGH</td>\n",
       "      <td>46.76527</td>\n",
       "      <td>-84.89710</td>\n",
       "      <td>Arrive St Marys River</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-28 19:27:17</th>\n",
       "      <td>70.0</td>\n",
       "      <td>8.15</td>\n",
       "      <td>1 days 00:38:51</td>\n",
       "      <td>46.10724</td>\n",
       "      <td>-84.01422</td>\n",
       "      <td>366971360.0</td>\n",
       "      <td>Depart St Marys River</td>\n",
       "      <td>8.3</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>JOHN G MUNSON</td>\n",
       "      <td>46.10983</td>\n",
       "      <td>-84.01380</td>\n",
       "      <td>Arrive St Marys River</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-31 14:15:16</th>\n",
       "      <td>70.0</td>\n",
       "      <td>8.03</td>\n",
       "      <td>2 days 05:39:00</td>\n",
       "      <td>46.76920</td>\n",
       "      <td>-84.90083</td>\n",
       "      <td>366971350.0</td>\n",
       "      <td>Depart St Marys River</td>\n",
       "      <td>6.7</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>PHILIP R CLARKE</td>\n",
       "      <td>46.76751</td>\n",
       "      <td>-84.89928</td>\n",
       "      <td>Arrive St Marys River</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-04-02 14:56:03</th>\n",
       "      <td>70.0</td>\n",
       "      <td>8.52</td>\n",
       "      <td>1 days 18:08:20</td>\n",
       "      <td>46.76925</td>\n",
       "      <td>-84.90090</td>\n",
       "      <td>367050550.0</td>\n",
       "      <td>Depart St Marys River</td>\n",
       "      <td>10.3</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>STEWART J CORT</td>\n",
       "      <td>46.76687</td>\n",
       "      <td>-84.89865</td>\n",
       "      <td>Arrive St Marys River</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-12-31 08:35:46</th>\n",
       "      <td>70.0</td>\n",
       "      <td>8.20</td>\n",
       "      <td>0 days 07:36:27</td>\n",
       "      <td>46.76726</td>\n",
       "      <td>-84.85710</td>\n",
       "      <td>366971360.0</td>\n",
       "      <td>Depart St Marys River</td>\n",
       "      <td>14.4</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>JOHN G MUNSON</td>\n",
       "      <td>46.75920</td>\n",
       "      <td>-84.85346</td>\n",
       "      <td>Arrive St Marys River</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-12-31 15:15:50</th>\n",
       "      <td>70.0</td>\n",
       "      <td>9.00</td>\n",
       "      <td>0 days 13:13:59</td>\n",
       "      <td>46.10797</td>\n",
       "      <td>-84.01491</td>\n",
       "      <td>316029000.0</td>\n",
       "      <td>Depart St Marys River</td>\n",
       "      <td>8.3</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>CSL NIAGARA</td>\n",
       "      <td>46.11064</td>\n",
       "      <td>-84.01456</td>\n",
       "      <td>Arrive St Marys River</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-12-31 17:38:25</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0 days 12:52:50</td>\n",
       "      <td>46.10703</td>\n",
       "      <td>-84.01505</td>\n",
       "      <td>366930730.0</td>\n",
       "      <td>Depart St Marys River</td>\n",
       "      <td>7.3</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>AMERICAN SPIRIT</td>\n",
       "      <td>46.11168</td>\n",
       "      <td>-84.01447</td>\n",
       "      <td>Arrive St Marys River</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-12-31 23:58:46</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0 days 12:38:19</td>\n",
       "      <td>46.46353</td>\n",
       "      <td>-84.57252</td>\n",
       "      <td>366971340.0</td>\n",
       "      <td>Depart St Marys River</td>\n",
       "      <td>0.1</td>\n",
       "      <td>at anchor</td>\n",
       "      <td>CASON J CALLAWAY</td>\n",
       "      <td>46.46355</td>\n",
       "      <td>-84.57252</td>\n",
       "      <td>Arrive St Marys River</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-12-31 23:58:52</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0 days 11:37:40</td>\n",
       "      <td>46.56716</td>\n",
       "      <td>-84.68026</td>\n",
       "      <td>316023339.0</td>\n",
       "      <td>Depart St Marys River</td>\n",
       "      <td>6.7</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>THUNDER BAY</td>\n",
       "      <td>46.56533</td>\n",
       "      <td>-84.67849</td>\n",
       "      <td>Arrive St Marys River</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>4604 rows × 13 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Cargo  Draft        Duration       LAT       LON  \\\n",
       "2015-03-26 18:36:48   75.0   9.77 2 days 13:31:44  46.76976 -84.90145   \n",
       "2015-03-26 18:46:36   70.0   8.51 2 days 06:05:28  46.76783 -84.89953   \n",
       "2015-03-28 19:27:17   70.0   8.15 1 days 00:38:51  46.10724 -84.01422   \n",
       "2015-03-31 14:15:16   70.0   8.03 2 days 05:39:00  46.76920 -84.90083   \n",
       "2015-04-02 14:56:03   70.0   8.52 1 days 18:08:20  46.76925 -84.90090   \n",
       "...                    ...    ...             ...       ...       ...   \n",
       "2017-12-31 08:35:46   70.0   8.20 0 days 07:36:27  46.76726 -84.85710   \n",
       "2017-12-31 15:15:50   70.0   9.00 0 days 13:13:59  46.10797 -84.01491   \n",
       "2017-12-31 17:38:25    NaN    NaN 0 days 12:52:50  46.10703 -84.01505   \n",
       "2017-12-31 23:58:46    NaN    NaN 0 days 12:38:19  46.46353 -84.57252   \n",
       "2017-12-31 23:58:52    NaN    NaN 0 days 11:37:40  46.56716 -84.68026   \n",
       "\n",
       "                            MMSI               Position   SOG  \\\n",
       "2015-03-26 18:36:48  366971370.0  Depart St Marys River   6.8   \n",
       "2015-03-26 18:46:36  366972030.0  Depart St Marys River   9.9   \n",
       "2015-03-28 19:27:17  366971360.0  Depart St Marys River   8.3   \n",
       "2015-03-31 14:15:16  366971350.0  Depart St Marys River   6.7   \n",
       "2015-04-02 14:56:03  367050550.0  Depart St Marys River  10.3   \n",
       "...                          ...                    ...   ...   \n",
       "2017-12-31 08:35:46  366971360.0  Depart St Marys River  14.4   \n",
       "2017-12-31 15:15:50  316029000.0  Depart St Marys River   8.3   \n",
       "2017-12-31 17:38:25  366930730.0  Depart St Marys River   7.3   \n",
       "2017-12-31 23:58:46  366971340.0  Depart St Marys River   0.1   \n",
       "2017-12-31 23:58:52  316023339.0  Depart St Marys River   6.7   \n",
       "\n",
       "                                     Status        VesselName  from_LAT  \\\n",
       "2015-03-26 18:36:48  under way using engine      EDWIN H GOTT  46.76789   \n",
       "2015-03-26 18:46:36  under way using engine      ROGER BLOUGH  46.76527   \n",
       "2015-03-28 19:27:17  under way using engine     JOHN G MUNSON  46.10983   \n",
       "2015-03-31 14:15:16  under way using engine   PHILIP R CLARKE  46.76751   \n",
       "2015-04-02 14:56:03  under way using engine    STEWART J CORT  46.76687   \n",
       "...                                     ...               ...       ...   \n",
       "2017-12-31 08:35:46  under way using engine     JOHN G MUNSON  46.75920   \n",
       "2017-12-31 15:15:50  under way using engine       CSL NIAGARA  46.11064   \n",
       "2017-12-31 17:38:25  under way using engine   AMERICAN SPIRIT  46.11168   \n",
       "2017-12-31 23:58:46               at anchor  CASON J CALLAWAY  46.46355   \n",
       "2017-12-31 23:58:52  under way using engine       THUNDER BAY  46.56533   \n",
       "\n",
       "                     from_LON                   From  \n",
       "2015-03-26 18:36:48 -84.89990  Arrive St Marys River  \n",
       "2015-03-26 18:46:36 -84.89710  Arrive St Marys River  \n",
       "2015-03-28 19:27:17 -84.01380  Arrive St Marys River  \n",
       "2015-03-31 14:15:16 -84.89928  Arrive St Marys River  \n",
       "2015-04-02 14:56:03 -84.89865  Arrive St Marys River  \n",
       "...                       ...                    ...  \n",
       "2017-12-31 08:35:46 -84.85346  Arrive St Marys River  \n",
       "2017-12-31 15:15:50 -84.01456  Arrive St Marys River  \n",
       "2017-12-31 17:38:25 -84.01447  Arrive St Marys River  \n",
       "2017-12-31 23:58:46 -84.57252  Arrive St Marys River  \n",
       "2017-12-31 23:58:52 -84.67849  Arrive St Marys River  \n",
       "\n",
       "[4604 rows x 13 columns]"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# We had some problems with the St Marys river Collection because it allowed min and max times to take place\n",
    "#anywhere along the channel.  We needed to place strict boundaries at the entry and exit to reduce noise.\n",
    "\n",
    "df=Trips.loc[(Trips.Position=='Depart St Marys River')&((Trips.LON<-84.57)|(Trips.LAT<46.12))&\n",
    "             ((Trips.from_LON<-84.57)|(Trips.from_LAT<46.12))]\n",
    "print(len(df))\n",
    "Trips.drop(Trips[Trips.Position=='Depart St Marys River'].index, inplace = True)\n",
    "Trips=Trips.append(df)\n",
    "Trips.loc[Trips.Position=='Depart St Marys River']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "736687\n",
      "AMERICAN CENTURY complete.\n",
      "AMERICAN INTEGRITY complete.\n",
      "AMERICAN SPIRIT complete.\n",
      "BUFFALO complete.\n",
      "BURNS HARBOR complete.\n",
      "CASON J CALLAWAY complete.\n",
      "CSL ASSINIBOINE complete.\n",
      "CSL LAURENTIEN complete.\n",
      "CSL NIAGARA complete.\n",
      "EDGAR B SPEER complete.\n",
      "EDWIN H GOTT complete.\n",
      "HERBERT C JACKSON complete.\n",
      "HON JAMES L OBERSTAR complete.\n",
      "JAMES R BARKER complete.\n",
      "JOHN G MUNSON complete.\n",
      "JOSEPH L BLOCK complete.\n",
      "JOYCE L VANENKEVORT complete.\n",
      "KAYE E BARKER complete.\n",
      "LEE A TREGURTHA complete.\n",
      "MESABI MINER complete.\n",
      "PHILIP R CLARKE complete.\n",
      "PRESQUE ISLE complete.\n",
      "ROGER BLOUGH complete.\n",
      "RT HON PAUL J MARTIN complete.\n",
      "SAM LAUD complete.\n",
      "STEWART J CORT complete.\n",
      "THUNDER BAY complete.\n",
      "TUG CLYDE S VANENKEVORT complete.\n",
      "VICTORY complete.\n",
      "WALTER J MCCARTHY JR complete.\n",
      "10073\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Cargo</th>\n",
       "      <th>Draft</th>\n",
       "      <th>Duration</th>\n",
       "      <th>LAT</th>\n",
       "      <th>LON</th>\n",
       "      <th>MMSI</th>\n",
       "      <th>Position</th>\n",
       "      <th>SOG</th>\n",
       "      <th>Status</th>\n",
       "      <th>VesselName</th>\n",
       "      <th>from_LAT</th>\n",
       "      <th>from_LON</th>\n",
       "      <th>From</th>\n",
       "      <th>Origin</th>\n",
       "      <th>Destination</th>\n",
       "      <th>Tons</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2015-03-25 12:53:46</th>\n",
       "      <td>Empty Barges</td>\n",
       "      <td>9.77</td>\n",
       "      <td>NaT</td>\n",
       "      <td>46.48675</td>\n",
       "      <td>-84.30111</td>\n",
       "      <td>366971370.0</td>\n",
       "      <td>Arrive Soo Locks</td>\n",
       "      <td>5.8</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>EDWIN H GOTT</td>\n",
       "      <td>46.48497</td>\n",
       "      <td>-84.29975</td>\n",
       "      <td>NaN</td>\n",
       "      <td>MILWAUKEE, WI</td>\n",
       "      <td>TWO HARBORS, MN</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-25 13:25:30</th>\n",
       "      <td>Empty Barges</td>\n",
       "      <td>8.51</td>\n",
       "      <td>NaT</td>\n",
       "      <td>46.48563</td>\n",
       "      <td>-84.30025</td>\n",
       "      <td>366972030.0</td>\n",
       "      <td>Arrive Soo Locks</td>\n",
       "      <td>4.0</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>ROGER BLOUGH</td>\n",
       "      <td>46.48460</td>\n",
       "      <td>-84.29943</td>\n",
       "      <td>NaN</td>\n",
       "      <td>ERIE, PA</td>\n",
       "      <td>TWO HARBORS, MN</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-26 01:40:38</th>\n",
       "      <td>Empty Barges</td>\n",
       "      <td>9.77</td>\n",
       "      <td>12:46:52</td>\n",
       "      <td>46.49918</td>\n",
       "      <td>-84.39629</td>\n",
       "      <td>366971370.0</td>\n",
       "      <td>Depart Soo Locks</td>\n",
       "      <td>9.6</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>EDWIN H GOTT</td>\n",
       "      <td>46.49979</td>\n",
       "      <td>-84.39251</td>\n",
       "      <td>Arrive Soo Locks</td>\n",
       "      <td>MILWAUKEE, WI</td>\n",
       "      <td>TWO HARBORS, MN</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-26 06:30:06</th>\n",
       "      <td>Empty Barges</td>\n",
       "      <td>8.51</td>\n",
       "      <td>17:04:36</td>\n",
       "      <td>46.49870</td>\n",
       "      <td>-84.39973</td>\n",
       "      <td>366972030.0</td>\n",
       "      <td>Depart Soo Locks</td>\n",
       "      <td>9.5</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>ROGER BLOUGH</td>\n",
       "      <td>46.49947</td>\n",
       "      <td>-84.39552</td>\n",
       "      <td>Arrive Soo Locks</td>\n",
       "      <td>ERIE, PA</td>\n",
       "      <td>TWO HARBORS, MN</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-27 22:10:47</th>\n",
       "      <td>Iron Ore</td>\n",
       "      <td>8.15</td>\n",
       "      <td>NaT</td>\n",
       "      <td>46.50180</td>\n",
       "      <td>-84.37771</td>\n",
       "      <td>366971360.0</td>\n",
       "      <td>Arrive Soo Locks</td>\n",
       "      <td>2.9</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>JOHN G MUNSON</td>\n",
       "      <td>46.49484</td>\n",
       "      <td>-84.41891</td>\n",
       "      <td>NaN</td>\n",
       "      <td>DULUTH, MN</td>\n",
       "      <td>GARY, IN</td>\n",
       "      <td>26837</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                            Cargo  Draft Duration       LAT       LON  \\\n",
       "2015-03-25 12:53:46  Empty Barges   9.77      NaT  46.48675 -84.30111   \n",
       "2015-03-25 13:25:30  Empty Barges   8.51      NaT  46.48563 -84.30025   \n",
       "2015-03-26 01:40:38  Empty Barges   9.77 12:46:52  46.49918 -84.39629   \n",
       "2015-03-26 06:30:06  Empty Barges   8.51 17:04:36  46.49870 -84.39973   \n",
       "2015-03-27 22:10:47      Iron Ore   8.15      NaT  46.50180 -84.37771   \n",
       "\n",
       "                            MMSI          Position  SOG  \\\n",
       "2015-03-25 12:53:46  366971370.0  Arrive Soo Locks  5.8   \n",
       "2015-03-25 13:25:30  366972030.0  Arrive Soo Locks  4.0   \n",
       "2015-03-26 01:40:38  366971370.0  Depart Soo Locks  9.6   \n",
       "2015-03-26 06:30:06  366972030.0  Depart Soo Locks  9.5   \n",
       "2015-03-27 22:10:47  366971360.0  Arrive Soo Locks  2.9   \n",
       "\n",
       "                                     Status     VesselName  from_LAT  \\\n",
       "2015-03-25 12:53:46  under way using engine   EDWIN H GOTT  46.48497   \n",
       "2015-03-25 13:25:30  under way using engine   ROGER BLOUGH  46.48460   \n",
       "2015-03-26 01:40:38  under way using engine   EDWIN H GOTT  46.49979   \n",
       "2015-03-26 06:30:06  under way using engine   ROGER BLOUGH  46.49947   \n",
       "2015-03-27 22:10:47  under way using engine  JOHN G MUNSON  46.49484   \n",
       "\n",
       "                     from_LON              From         Origin  \\\n",
       "2015-03-25 12:53:46 -84.29975               NaN  MILWAUKEE, WI   \n",
       "2015-03-25 13:25:30 -84.29943               NaN       ERIE, PA   \n",
       "2015-03-26 01:40:38 -84.39251  Arrive Soo Locks  MILWAUKEE, WI   \n",
       "2015-03-26 06:30:06 -84.39552  Arrive Soo Locks       ERIE, PA   \n",
       "2015-03-27 22:10:47 -84.41891               NaN     DULUTH, MN   \n",
       "\n",
       "                         Destination   Tons  \n",
       "2015-03-25 12:53:46  TWO HARBORS, MN      0  \n",
       "2015-03-25 13:25:30  TWO HARBORS, MN      0  \n",
       "2015-03-26 01:40:38  TWO HARBORS, MN      0  \n",
       "2015-03-26 06:30:06  TWO HARBORS, MN      0  \n",
       "2015-03-27 22:10:47         GARY, IN  26837  "
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#This algorithm collects start and complete positions for each ship in the Soo Locks \n",
    "#3-4) Arrival and departure from the Soo Locks in Sault Ste Marie\n",
    "\n",
    "df=Iron\n",
    "W,E,N,S=features.loc['St Marys River']\n",
    "W,E=-84.4,-84.3 #placing stricter boundaries around the Soo Locks\n",
    "df1=df.loc[(df.LAT<=N)&(df.LAT>=S)&(df.LON<=E)&(df.LON>=W)] #subset of data for single geographic feature\n",
    "\n",
    "print(len(df1))\n",
    "\n",
    "Locks=pd.DataFrame() #empty data frame to collect data points\n",
    "for vessel in vessels:\n",
    "    df2 = df1.loc[(df1.VesselName==vessel)]#&(df1.SOG<0.2)]\n",
    "    \n",
    "    Lock=pd.DataFrame() #empty data frame to collect data points\n",
    "\n",
    "    datelist=np.unique(df2.index.date)\n",
    "    dates=datelist[(datelist-np.roll(datelist,1))!=pd.Timedelta(days=1)]\n",
    "\n",
    "    for date in dates:\n",
    "        df3=df2.loc[df2.index.date==date]\n",
    "        add=df3.loc[min(df3.index)]\n",
    "        add['Position']='Arrive Soo Locks'\n",
    "        Lock=Lock.append(add)\n",
    "\n",
    "    dates=datelist[(datelist-np.roll(datelist,-1))!=pd.Timedelta(days=-1)]\n",
    "    for date in dates:\n",
    "        df3=df2.loc[df2.index.date==date]\n",
    "        add=df3.loc[max(df3.index)]\n",
    "        add['Position']='Depart Soo Locks'\n",
    "        Lock=Lock.append(add)\n",
    "\n",
    "\n",
    "    Lock= Lock.sort_index()\n",
    "    Lock['Duration']=Lock.index.to_series().diff()\n",
    "    Lock['From']=Lock.Position.shift()\n",
    "    lpms= LPMS.loc[(LPMS.VesselName==vessel)]\n",
    "    dates=np.unique(lpms.index.date)  \n",
    "    for date in dates:\n",
    "        df4=Lock.loc[(Lock.index.date>=(date-dt.timedelta(days=1)))&(Lock.index.date<=(date+dt.timedelta(days=1)))]\n",
    "        df5=lpms.loc[lpms.index.date==date].reset_index()\n",
    "        df4['Origin']=df5.loc[0,'Origin']\n",
    "        df4['Destination']=df5.loc[0,'Destination']\n",
    "        df4['Cargo']=df5.loc[0,'Cargo']\n",
    "        df4['Tons']=df5.loc[0,'Tons']\n",
    "        Locks=Locks.append(df4).sort_index()\n",
    "    print(f'{vessel} complete.')\n",
    "\n",
    "Locks = Locks.sort_index()\n",
    "print(len(Locks))\n",
    "Locks.head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "42021\n",
      "42021\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Cargo</th>\n",
       "      <th>Draft</th>\n",
       "      <th>Duration</th>\n",
       "      <th>LAT</th>\n",
       "      <th>LON</th>\n",
       "      <th>MMSI</th>\n",
       "      <th>Position</th>\n",
       "      <th>SOG</th>\n",
       "      <th>Status</th>\n",
       "      <th>VesselName</th>\n",
       "      <th>from_LAT</th>\n",
       "      <th>from_LON</th>\n",
       "      <th>From</th>\n",
       "      <th>Origin</th>\n",
       "      <th>Destination</th>\n",
       "      <th>Tons</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2015-03-01 03:14:00</th>\n",
       "      <td>NaN</td>\n",
       "      <td>8.03</td>\n",
       "      <td>0.0</td>\n",
       "      <td>41.69872</td>\n",
       "      <td>-83.45046</td>\n",
       "      <td>366971350.0</td>\n",
       "      <td>Depart Toledo</td>\n",
       "      <td>0.1</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>PHILIP R CLARKE</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Arrive Toledo</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-01 03:14:00</th>\n",
       "      <td>NaN</td>\n",
       "      <td>8.03</td>\n",
       "      <td>NaN</td>\n",
       "      <td>41.69872</td>\n",
       "      <td>-83.45046</td>\n",
       "      <td>366971350.0</td>\n",
       "      <td>Arrive Toledo</td>\n",
       "      <td>0.1</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>PHILIP R CLARKE</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-02 16:50:01</th>\n",
       "      <td>NaN</td>\n",
       "      <td>9.01</td>\n",
       "      <td>NaN</td>\n",
       "      <td>42.88318</td>\n",
       "      <td>-79.24848</td>\n",
       "      <td>316001635.0</td>\n",
       "      <td>Arrive Welland Canal</td>\n",
       "      <td>0.0</td>\n",
       "      <td>under way sailing</td>\n",
       "      <td>RT HON PAUL J MARTIN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-02 19:26:23</th>\n",
       "      <td>NaN</td>\n",
       "      <td>9.01</td>\n",
       "      <td>156.0</td>\n",
       "      <td>42.88314</td>\n",
       "      <td>-79.24848</td>\n",
       "      <td>316001635.0</td>\n",
       "      <td>Depart Welland Canal</td>\n",
       "      <td>0.0</td>\n",
       "      <td>under way sailing</td>\n",
       "      <td>RT HON PAUL J MARTIN</td>\n",
       "      <td>42.88314</td>\n",
       "      <td>-79.24848</td>\n",
       "      <td>Arrive Welland Canal</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-03-03 11:28:41</th>\n",
       "      <td>NaN</td>\n",
       "      <td>8.03</td>\n",
       "      <td>3374.0</td>\n",
       "      <td>41.69873</td>\n",
       "      <td>-83.45055</td>\n",
       "      <td>366971350.0</td>\n",
       "      <td>Arrive Toledo</td>\n",
       "      <td>0.1</td>\n",
       "      <td>under way using engine</td>\n",
       "      <td>PHILIP R CLARKE</td>\n",
       "      <td>41.69872</td>\n",
       "      <td>-83.45046</td>\n",
       "      <td>Depart Toledo</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    Cargo  Draft  Duration       LAT       LON         MMSI  \\\n",
       "2015-03-01 03:14:00   NaN   8.03       0.0  41.69872 -83.45046  366971350.0   \n",
       "2015-03-01 03:14:00   NaN   8.03       NaN  41.69872 -83.45046  366971350.0   \n",
       "2015-03-02 16:50:01   NaN   9.01       NaN  42.88318 -79.24848  316001635.0   \n",
       "2015-03-02 19:26:23   NaN   9.01     156.0  42.88314 -79.24848  316001635.0   \n",
       "2015-03-03 11:28:41   NaN   8.03    3374.0  41.69873 -83.45055  366971350.0   \n",
       "\n",
       "                                 Position  SOG                  Status  \\\n",
       "2015-03-01 03:14:00         Depart Toledo  0.1  under way using engine   \n",
       "2015-03-01 03:14:00         Arrive Toledo  0.1  under way using engine   \n",
       "2015-03-02 16:50:01  Arrive Welland Canal  0.0       under way sailing   \n",
       "2015-03-02 19:26:23  Depart Welland Canal  0.0       under way sailing   \n",
       "2015-03-03 11:28:41         Arrive Toledo  0.1  under way using engine   \n",
       "\n",
       "                               VesselName  from_LAT  from_LON  \\\n",
       "2015-03-01 03:14:00       PHILIP R CLARKE       NaN       NaN   \n",
       "2015-03-01 03:14:00       PHILIP R CLARKE       NaN       NaN   \n",
       "2015-03-02 16:50:01  RT HON PAUL J MARTIN       NaN       NaN   \n",
       "2015-03-02 19:26:23  RT HON PAUL J MARTIN  42.88314 -79.24848   \n",
       "2015-03-03 11:28:41       PHILIP R CLARKE  41.69872 -83.45046   \n",
       "\n",
       "                                     From Origin Destination  Tons  \n",
       "2015-03-01 03:14:00         Arrive Toledo    NaN         NaN   NaN  \n",
       "2015-03-01 03:14:00                   NaN    NaN         NaN   NaN  \n",
       "2015-03-02 16:50:01                   NaN    NaN         NaN   NaN  \n",
       "2015-03-02 19:26:23  Arrive Welland Canal    NaN         NaN   NaN  \n",
       "2015-03-03 11:28:41         Depart Toledo    NaN         NaN   NaN  "
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Trips.Cargo=np.nan\n",
    "Combo=Locks.append(Trips).sort_index()\n",
    "print(len(Combo))\n",
    "Filled=pd.DataFrame()\n",
    "for vessel in vessels:\n",
    "    df=Combo.loc[Combo.VesselName==vessel]\n",
    "    dates=np.unique(df.index.date)\n",
    "    date=min(dates)\n",
    "    #df=df.ffill(limit=3).bfill(limit=3)\n",
    "    #df=df.ffill(limit=2).bfill()\n",
    "    Filled=Filled.append(df).sort_index()\n",
    "\n",
    "print(len(Filled))\n",
    "Filled.Duration=Filled.Duration.astype('timedelta64[m]')\n",
    "Filled.to_csv('Merged_Data_new.csv')\n",
    "Filled.head()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}