Deliverables/Documentation

Website

https://yanni-ma.github.io/EPI-Project/

Project Deliverables

• Project Code - Full Repo:
  GitHub Repository

• Project Code - Tarball:
  Tarball

• Project Poster:
  Google Slides Poster

• Project Final Report:
  Final Report

Political SSEs Impact on COVID-19 Case Count Analysis and Forecasting

This repository contains Python scripts for analyzing and forecasting COVID-19 case counts using advanced statistical models like SARIMAX and SEIR. The project incorporates event-based data (valence, size_mean) to study the potential impacts of political events on case trends.

important files

• SEIR_with_counties.py: Simulates the spread of infections using the SEIR model, incorporating county-level data
• daily_arima.py: Fits ARIMA models with smoothing to forecast case counts
• daily_arima_without_smoothing.py: Fits ARIMA models without smoothing for comparison
• data_preprocessing.py: Contains helper functions for loading and preprocessing datasets
• main.py: File for analyzing and visualizing case trends with basic forecasting
• minimize_RMSE_for_smoothed_ARIMA.py: Grid search to minimize RMSE for ARIMA models with smoothing
• minimize_RMSE_smoothed_no_exog.py: Grid search to minimize RMSE for ARIMA models without exogenous variables
• modified_main_to_detect_spikes.py: Script for detecting spikes in case counts from preprocessed data
• unweighted_arima_valence_case_counts_projection.py: ARIMA forecasting considering valence without weighting by event size
• weighted_arima_valence_case_counts_projection.py: ARIMA forecasting considering valence, weighted by event size
• measure_arima_performance.py: Compares performance metrics (e.g., RMSE) of different ARIMA models

Datasets

The datasets are provided in the dataset folder. They are sourced from NYT (https://github.com/nytimes/covid-19-data) and CCC (https://github.com/nonviolent-action-lab/crowd-counting-consortium)

Usage

To run any Python script, use the following command in your terminal while cd'd into the SRC folder:

python <script_name>.py