Bridging the Gap: Enhancing COVID-19 Epidemic Forecasting by Integrating Factors like Vaccination Rates, Mobility, Stringency, Socio-Economic Indicators, and Health Metrics into Time Series Models

Contributors

• Mehul Rastogi (mehulrastogi at gatech dot edu)
• Akshat Karwa (akshatkarwa at gatech dot edu)

Website

More information about the project can be found on this website - Project Website.

Setup

pip install prophet scikit-learn statsmodels pandas matplotlib sktime scalecast

Package Structure

• All the code files are in the path: src/.
• The raw and cleaned data is in data/.
• All the plots are in plots/.

Data Cleaning

cd src
python3 source.py

All the data cleaning and pre-processing, changes can be seen in /data

SARIMA and Mobility Graphs

cd src
python3 mobility_model.py

This will run the SARIMA model, return the metrics and update all the plots in plots/SARIMA. This will also generate the mobility graphs in plots/mobility.

FAQ - If you're interested - Add a new case case5 array on line 105 to try out the computation with different sets of regressor/exogenous variables. Also, add that array in the for loop on line 107.

Vaccination Indicator

cd src
python3 vaccination_indicator.py

This will generate all the plots in plots/vaccination

Meta Prophet Model

cd src
python3 meta_prophet_model.py

Simulate the Meta Prophet model. Also, generate graphs and metrics for the same. Plots to be found in plots/Prophet.

FAQ - If you're interested - Add a new case case4 array on line 51 to try out the computation with different sets of regressor/exogenous variables. Also, add that array in the for loop on line 53.

TBATS Model

cd src
python3 tbats_model.py

Simulate the TBATS model. Also, generate plots that will be saved in plots/tbats.

SI Model

cd src
python3 si_model.py

Simulate the SI model after data processing. Also, generate plots that will be saved in plots/SI_model.

RNN and LSTM models

cd src
python3 run_RNN_LSTM.py

This file runs the RNN and LSTM models and generates plots that are saved in plots/RNN. Layer Functions are in LSTM_RNN.py. Also generates metrics that are in src/results.xlsx.

Documentation & Reports

The following reports can be found in the doc/ directory:

• Project Proposal
• Project Milestone
• Final Project Report