Corelation of weather parameters with COVID-19 spread and mortality rate
In this work, we explore the corelation of weather parameters with COVID 19 spread and mortality rate.
Here we are motivated to study the effect of weather conditions, area density and infection (case count) due to COVID-19, and mortality rate. We have used several machine learning models for exploring these relationships. We have used publicly available datasets from Kaggle for validations. The performance of the Machine learning regression models is measured using standard performance metrics.
Goals And Challenges
In this project, we first estimate the relation between different weather parameters and covid-19 spread and convert it to a model to forecast the spread of the virus at a time according to dependency upon other weather variables. This model takes a very correlational view of the data and explores the inter-zone differences in the virus spread. The seasonal nature of flu spread drives our modeling hypothesis, and associated weather variances inform our thinking on this hypothesis.
We are using weather differences across different zones/times to understand the correlations.However, lack of availabile long time series data at this point, is limiting our study to exploratory analysis.
Weather parameter : Temperature
Temperature VS covid cases
(color bar denotes temperature and size of bubble denotes the number of covid cases)
Graph for the dependency of covid confirmed case on weather parameters and other different factors using decision tree model
Feature importance graph for the dependency of covid deaths on weather parameters and other different factors using decision tree model
Heatmap plot for correlation between different variables used.
Copyright© 2020-21. CARE Inc. All Rights Reserved.