This is a github repo of my contributions to a group project.
Utilizing weather, geographic, and mosquito collection data from Chicago, my team was able to build a model that predicted where and when mosquito populations would increase and if they would be likely carriers of the West Nile Virus. We then did a cost-benefit analysis on the effects of spraying pesticide to control the mosquito population and thus the spread of the disease. Our model was able to predict 81% of actual cases of WNV based on our test data. We also score a .77355 on the Roc Auc score (used to find the winner on Kaggle).
You can find the presentation we gave here: https://docs.google.com/presentation/d/1jYtqXynF6ohSalfJJv15OI5EVDpOy20XyFamecFL8Eo/edit?usp=sharing
My contributions include:
- Exploritory Data Analysis
- Geographical Visualization
- Exploring the effects of applying spatial lags to our data using Python Library pysal
Other group members were:
Dale Wahl - https://github.com/dale-wahl
Diego Rodriguez - https://rodriguezda.github.io/
Kyle Santana - http://kylesantana.com
Matt Bollinger - https://github.com/mlybollinger