Chicago/sdp-feature-fetcher: Construct modeling features based on business.json files and Plenar.io

This package is part of the Smart Data Platform developed by the City of Chicago. The Feature Fetcher is designed to ingest data produced by other components of the Smart Data Platform (SDP) and automatically create variables that can then be used in downstream modeling process. Starting with a business.json file which follows the specification of the SDP business.json format (https://github.com/Chicago/sdp-business.json) the Feature Fetcher provides tools to form the JSON into a table with fields that are calculated based on the inspection and business information contained in the data. Fields such as "number of inspections in past x days" can be automatically calculated, as well as a number of other fields generally used in the model process. This core table can be joined to other event based data that is calculated based on the inspection dates given in the table. Event based data is calculated based on a Gaussian Kernel Density Estimate.

install.packages("remotes")
remotes::install_github("Chicago/sdp-feature-fetcher")