Urban Data Watch

A tool for democritizing data-grounded hypothesis testing and exploration

About The Project

The project aims to build a tool for running data-grounded examinations for generating and verifying hypotheses one may have concerning a city.

Data Watch Methodology

The tool borrows from a well-known GIS data analytics model called Geographically Weighted Regression. It is just a slight modification of the linear regression model that considers spatially varying relationships. In general, regression analysis is useful to inspect how a variable is relevant to fitting the data to the dependent variable.

Paper

Regression Result

News Gap View

Using CLIFF, we can map news articles to corresponding census tracts to see how well a region is covered based on the crime terms mentioned. This can be potential extension to visualize analyze the relevance of a candidate attribute and corresponding news coverage volume (or lack of it).