The number of confirmed novel coronavirus cases in India has reached close to 750 cases as of 27th March 2020. According to the health ministry, Covid-19, which has globally infected over 1.5 lakh people, is likely to cause more shutdowns across the country as more states rush to stop the virus from spreading. While 13 people who were infected by the virus have recovered in India, the global situation still paints a grim picture. The death toll due to new coronavirus has shot up sharply in Italy, which has now become the country worst affected by the virus outbreak after China. Now US cases have also risen drastically. Besides taking precautions to prevent the spread of the virus domestically, India is also actively rescuing its nationals stuck in the high-affected zone as well as trying to help its neighbours in this warlike effort.
The WHO, Governments of the world have made open source several datasets related to COVID-19 for the Scientific community to study, analyse and may be come up with possible measures to prevent or stop the spread of the same
2. The objective of the current study
In this study, we compare the rise of COVID Cases for India and US with a few selected countries. We further aim to compare the spread of COVID-19 in India and US and tries to predict the possible cases until 1st April using data up to 25th March. We try to see how using a very simple univariate data available from John Hopkins University how accurately one can predict e rise with simple mathematical modelling.
3. Data Sources / Data Acknowledgements
John Hopkins University
4. GITHUB Source code
5. Visualization studies: India with predictions
- COVID-19 Rise compared India vs Iran, Italy, Korea and Spain
- COVID-19 Deaths compared India vs Iran, Italy, Korea and Spain
- COVID-19 Predictions in India till 1st April 2020
- COVID-19 Predictions vs Actuals in India: Lock Down Effect
COVID-19 rise compared India vs Iran, Italy, Korea and Spain
COVID-19 deaths compared India vs Iran, Italy, Korea and Spain
COVID-19 predictions till 1st April in India 2020
COVID-19 predictions vs Actuals – Lock Down Effect
6. Visualization studies: US with predictions
- COVID-19 Spread compared US vs EU
- EU vs US nos of days with greater than 100 cases
- Fitting 3rddegree polynomial to predict US COVID RISE
COVID-19 spread compared US vs EU
EU vs US: number of days with greater than 100 cases
Fitting 3rd Degree polynomial to predict US COVID-19 RISE
7. Mathematical Modelling Approach
In order to predict the COVID-19 rise for India and US, A simple but effective curve fitting has been employed. The univariate public source data is filtered to create a numeric series starting from the day where 100 cases or more have been reached. That as our start point we then fit a 3rd-degree polynomial till the data up to 25th March. Once we have our polynomial we then use it to predict a forward-looking window of 7 days, in this case till 1st April. Special comparison for Lockdown as in India has not been modelled.
With experts, the world over trying to come up with sophisticated models to predict COVID-19 rise quickly, this paper attempts to address the problem using a very simple numerical methods approach based on near history. The results look promising for the immediate future. As lockdowns start taking shape and the weather changes as well more sophisticated models have to employed going hence to reach accurate levels of prediction specially for large countries like US and India
Author and Acknowledgments
Dr. Anish Roy Chowdhury is currently an Industry Data Science Leader at Brillio, a Leading Digital Services Organization. In previous roles he was with ABInBev as a Data Science Research lead working in areas of Assortment Optimization, Reinforcement Learning to name a few, He also led several machine learning projects in areas of Credit Risk, Logistics and Sales forecasting. In his stint with HP Supply Chain Analytics he developed data Quality solutions for logistics projects and worked on building statistical models to predict spares part demands for large format printers. Prior to HP, he has 6 years of Work Experience in the IT sector as a DataBase Programmer. During his stint in IT he has worked for Credit Card Fraud Detection among other Analytics related Projects. He has a PhD in Mechanical Engineering (IISc Bangalore). He also holds an MS degree in Mechanical Engineering from Louisiana State University, USA. He did his undergraduate studies from NIT Durgapur with published research in GA- Fuzzy Logic applications to Medical diagnostics.
Dr. Anish is also a highly acclaimed public speaker with numerous best presentation awards from National and international conferences and has also conducted several workshops in Academic institutes on R programming and MATLAB. He also has several academic publications to his credit and is also a Chapter Co-Author for a Springer Publication and a Oxford University Press, best selling publication in MATLAB.
The author wishes to acknowledge Dr. Jacob Minz, Senior Research Manager at Synopsis and one of India’s earliest AI Evangelists, for the constant motivation and support to have this work attempted at a quick turnaround and short notice. The author also wishes to thank Parul Pandey, A data Science Evangelist currently with H2o.ai whose initial work in Kaggle was an inspiration to take up this study.