Maximizing Insight from Modern Economic Analysis
dc.contributor.author | Antenucci, Dolan | |
dc.date.accessioned | 2018-06-07T17:46:12Z | |
dc.date.available | NO_RESTRICTION | |
dc.date.available | 2018-06-07T17:46:12Z | |
dc.date.issued | 2018 | |
dc.date.submitted | ||
dc.identifier.uri | https://hdl.handle.net/2027.42/144007 | |
dc.description.abstract | The last decade has seen a growing trend of economists exploring how to extract different economic insight from "big data" sources such as the Web. As economists move towards this model of analysis, their traditional workflow starts to become infeasible. The amount of noisy data from which to draw insights presents data management challenges for economists and limits their ability to discover meaningful information. This leads to economists needing to invest a great deal of energy in training to be data scientists (a catch-all role that has grown to describe the usage of statistics, data mining, and data management in the big data age), with little time being spent on applying their domain knowledge to the problem at hand. We envision an ideal workflow that generates accurate and reliable results, where results are generated in near-interactive time, and systems handle the "heavy lifting" required for working with big data. This dissertation presents several systems and methodologies that bring economists closer to this ideal workflow, helping them address many of the challenges faced in transitioning to working with big data sources like the Web. To help users generate accurate and reliable results, we present approaches to identifying relevant predictors in nowcasting applications, as well as methods for identifying potentially invalid nowcasting models and their inputs. We show how a streamlined workflow, combined with pruning and shared computation, can help handle the heavy lifting of big data analysis, allowing users to generate results in near-interactive time. We also present a novel user model and architecture for helping users avoid undesirable bias when doing data preparation: users interactively define constraints for transformation code and the data that the code produces, and an explain-and-repair system satisfies these constraints as best it can, also providing an explanation for any problems along the way. These systems combined represent a unified effort to streamline the transition for economists to this new big data workflow. | |
dc.language.iso | en_US | |
dc.subject | economic big data analysis | |
dc.title | Maximizing Insight from Modern Economic Analysis | |
dc.type | Thesis | en_US |
dc.description.thesisdegreename | PhD | en_US |
dc.description.thesisdegreediscipline | Computer Science & Engineering | |
dc.description.thesisdegreegrantor | University of Michigan, Horace H. Rackham School of Graduate Studies | |
dc.contributor.committeemember | Cafarella, Michael John | |
dc.contributor.committeemember | Shapiro, Matthew D | |
dc.contributor.committeemember | Jagadish, Hosagrahar V | |
dc.contributor.committeemember | Koutra, Danai | |
dc.contributor.committeemember | Mozafari, Barzan | |
dc.subject.hlbsecondlevel | Computer Science | |
dc.subject.hlbtoplevel | Engineering | |
dc.description.bitstreamurl | https://deepblue.lib.umich.edu/bitstream/2027.42/144007/1/dol_1.pdf | |
dc.identifier.orcid | 0000-0001-9208-6961 | |
dc.identifier.name-orcid | Antenucci, Dolan; 0000-0001-9208-6961 | en_US |
dc.owningcollname | Dissertations and Theses (Ph.D. and Master's) |
Files in this item
Remediation of Harmful Language
The University of Michigan Library aims to describe library materials in a way that respects the people and communities who create, use, and are represented in our collections. Report harmful or offensive language in catalog records, finding aids, or elsewhere in our collections anonymously through our metadata feedback form. More information at Remediation of Harmful Language.
Accessibility
If you are unable to use this file in its current format, please select the Contact Us link and we can modify it to make it more accessible to you.