Abstract
Healthcare indices are used to evaluate the overall accessibility of public health resources for a given region and have been found to be especially useful for predicting post-treatment outcomes for a variety of diseases. Heart failure (HF) patients’ outcome after receiving treatment, for instance, has been discovered to be strongly correlated with the patient’s residing neighborhood, a variable yet to be incorporated into the process of health indices’ generation. Based on preliminary research that shows census-level Twitter data can be utilized to capture neighborhood impact, we introduce a visual analytic system that enables the iterative refinement of this new data. Specifically, the system enables machine learning experts, healthcare professionals, and policymakers to (1) preprocess tweets retrieved by specified keywords, e.g., eliminating URLs, stemming words, etc., (2) extract keywords and their corresponding embeddings, and (3) customize, inspect, and refine a topic model through interactive clustering. Ultimately, data produced by this system, along with other data sources, allows for the refinement of a more environment-reflective healthcare index.
Table of Contents
Introduction Background and Related Work System Design User Study Discussion Conclusion
About this Honors Thesis
Rights statement
- Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.
| School |
|
|---|
| Department |
|
|---|
| Degree |
|
|---|
| Submission |
|
|---|
| Language |
|
|---|
| Research Field |
|
|---|
| 关键词 |
|
|---|
| Committee Chair / Thesis Advisor |
|
|---|
| Committee Members |
|
|---|