This repository contains analysis of COVID-19 Non-pharmaceutical Interventions events data recorded in the WNTRAC dataset. Analysis has done for a few African countries, but they can be extended to other countries as the dataset contains worldwide events. Four notebooks are included in this repository
The notebook wntrac_descriptive_statistics.ipynb includes analysis of
WNTRAC dataset
by comparing Non-pharmaceutical Interventions implemented worldwide against
those implemented by Uganda, DRC Congo, Senegal & Nigeria.
Earlier version of this analysis were presented in our
paper in figures 6-12 generation code.
The notebook wntrac_correlation_analysis.ipynb includes correlation analysis of
NPIs, beta transmission parameter and COVID-19 cases data for 6 African countries,
including Uganda, DRC Congo, Senegal, Nigeria, Kenya & South Africa.
The notebook wntrac_adherence_index.ipynb describes and shows derivations of WNTRAC adherence index
by weighting the sum of
Oxford Stringency Index
(SI) and Compliance Score (CS).
This tutorial pertains to training and evaluating an interrupted time series model in Python using the open-source Prophet library by Facebook Inc. (https://facebook.github.io/prophet/)
The notebooks are standalone and can be run independently. We recommend running them in the IBM Cloud PAK for Data and follow the instructions there.
The notebooks can also be run locally by ensuring that a local python3
version is installed in the machine, and pip3 install .. all the missing libraries
that have been imported in the notebooks.