COVID-19 Risk Estimation for L.A. County using a Bayesian Time-varying SIR-model

This repository presents a rigorous hybrid model-and-data-driven Bayesian approach to risk scoring that yields a simplified color-coded risk level for each community. The risk score corresponds to the probability of someone currently healthy getting infected with COVID-19 in the near future.

Data Source

This project used CoVID-19 case data from LA County Dep. of Public Health, which is also available in the following repository: Data.

Instructions for running the following codes

Note: These files are developed and tested in Python3.6.

The following packages are required to run the codes:

• jupyter noteboook
• numpy
• json
• re
• matplotlib
• itertools
• pandas
• scipy
• gekko
• IPython
• datetime

These packages can be installed by issuing the following command "pip install name_of_the_package".

Running estimation_prediction_for_RiskScore_and_R.ipynb

Required input files: “Covid-19.csv”, “lacounty_covid.json”, “population.json”, “Covid-19-density.csv”

Output: -Computing and displaying the risk score as well as Rt for both the entire LA county and its communities. For simplicity, we illustrate it for the following 4 communities, namely 'West Hollywood', 'East Los Angeles', 'Castaic', and 'San Pedro'.

Parameters:

• number_of_days_passed_from_16th: #days passed March 16,2020 (e.g. it is 84 until June 7, 2020)
• show_Risk : True if showing risk score , otherwise False to show Rt and its confidence interval
• Whole_LAcounty: True to plot for entire LA county, and False to plot for the 4 communities
• moving_average_days: used for smoothing the curves

Running estimation_R_for_heatmap_histogram_of_RiskScore_and_R.ipynb

Required input files: “Covid-19.csv”, “lacounty_covid.json”, “population.json”, “Covid-19-density.csv”

Output:

• Computing and displaying the histogram of risk scores as well as Rt across all communities
• Generating a csv file for showing a heatmap of risk scores

Parameters:

• number_of_days_passed_from_16th: #days passed March 16,2020 (e.g. it is 84 until June 7, 2020)
• moving_average_days: used for smoothing the curves

Running generate_heatmap.py

Required input files: “Covid-19.csv”, “lacounty_covid.json”, “population.json”, “Covid-19-density.csv”, "la.shp"

Output:

• Generate heatmaps for each day

Execution command:

• python3.x generate_heatmap.py

Questions

For any questions about this project, please contact Prof. Bhaskar Krishnamachari ([email protected]).