This project demonstrates the capability to predict the overall health score of restaurants in each zip code of New York City by using 311 complaints data, DOHMH inspection data, and social media data from Yelp API.
To install the dependencies, run:
pip install yelpapi.
Download the datasets specified in the input folder.
Run the code to parse 311 and DOHMH inspection data:
python 311_parser.py
python DOHMH_parser.py
Run the code to get the average scores and review counts of restaurants and bars
using Yelp API:
python yelp_parser.py restaurants
python yelp_parser.py bars
Run the code to merge all parsed data from different sources together:
python merger.py
To see the regression result on the merged dataset merged.csv:
PCA --> run the notebook analysis_PCA.ipynb,
Lasso --> run the notebook analysis_Lasso.ipynb,
Random Forest --> run the notebook analysis_RandomForest.ipynb
All parsed datasets, plus a json file specifying the weights for the Lasso model,
are in the output folder.
All plots and maps are in the figures folder.