aprilgriffin/west_nile_modeling_mapping

Clean version of identifying_wnv_locations

IDENTIFYING WNV+ HOTSPOTS IN CHICAGO

Table of Contents:

1. Exploratory Data Analysis (EDA) Notebook
2. Geopandas Geo-mapping Notebook
3. Modeling Notebook
4. Data
   • Chicago Geodata (for mapping)
   • Kaggle Datasets
   • CSV file of locations from best model
5. Slide Deck (.pdf)

This repository contains work from a project completed by April Griffin, Suzanne McGann, and Ryan Stewart. See this link for the Kaggle data and competiton details: https://www.kaggle.com/c/predict-west-nile-virus

Modeling inclued RandomForest, XGBoost, AdaBoost and a Logistic Regression Pipeline on both dummied and purely numerical data. The AdaBoostClassifier on purely numerical data was our best performaing submission on Kaggle.

Cost-benefit analysis was discussed in a live presentation, and the slide deck (.pdf) is included in this repo.