PROTOTYPE: Using ML to Construct a Temporally-Evolving Heat Vulnerability Index (HVI) for NYC

Relevant Scripts

aggregateData.R

Inputs: Datasets from data folder

Outputs: allData.csv, data_NAandSteamOmitted.csv

Goal: take heat vulnerability-relevant datasets and create a .csv file with relevant variables across time and ZIP code.

ML Tree Stuff.Rmd

Goal: prepare tree counts and 311 calls datasets for use in aggregateData.R

Final_Prototype_Pipeline.ipynb

Inputs: data_NAandSteamOmitted.csv

Outputs: Decision tree, random forest, and XGBoost models for temporal HVI

Goal: Use machine learning to create baseline models with three different methods to predict HVI temporally for each ZIP code. Ground-truth these models against the given HVI from the NYC government and see if HVI changes temporally.

Other Components

environment.yml

Necessary dependencies for .ipynb files

data folder

Stores .csv files for all used datasets