Skip to content

LindsayMoir/covid

master
Switch branches/tags
Code

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Data Science Blog Post - Covid-19, Time To Get Out?

May 5, 2020

Table of Contents

Installation

The code runs with Python version 3.6.3. There are a large number of libraries required to run this notebook. Those are for ARIMA, BOKEH, and "regular libraries". Please go to the # import libraries cell which is near the top of the notebook for the definitive list.

Project Motivation

It can be difficult to get accurate and time sensitive information on how countries are doing at present during the SARS-2 pandemic (COVID-19 is the illness, the pandemic is the virus). Exploring factors potentially related to outcomes is important as a means to consider further outcome prediction. Key factors are likely related to government and public health leadership and the potential for actual public adoption of mitigation strategies for the virus. In this post, I answer 4 questions that should provide some insight into this very important issue.

  • Q1: Who are the best and worst countries dealing with Covid-19?
  • Q2: Which countries have flattened or are flattening the curve?
  • Q3: Can I see a global geographic representation of infections?
  • Q4: What is the projected global mortality by December 31, 2020?

File Descriptions

The following data files (available in this repository) are required.

Results

The results are published in this blog post. A quick summary is this.

  • 24 of 28 countries (that met the cutoffs) have a flattening of the curve presently. That is VERY encouraging.
  • It is surprising to see rich countries struggle with this virus. There is a saying that democracies get the type of leadership that they deserve. Hopefully this is a wake up call that electing incompetent politicians with massive character flaws to the highest office in the land can and does cause death. This time those deaths are being counted. They are not hidden in Syria, Afghanistan, etc. They are right in the home countries that have weak leadership.
  • Including the Median Per Capita Income and the Transparency International score was helpful in terms of creating cutoffs that made sense.
  • One sobering thought here is that there are 203 unique Alpha_3 codes in this dataset. That corresponds to 203 Country_Regions. We are only looking at 28 of them. This is because of the concern over the honesty of the data. In otherwords the challenge that we are looking at here is only the tip of the iceberg.

Deploy

Update the John Hopkins, Covid-19 data https://github.com/CSSEGISandData/COVID-19 by refreshing your local copy of the GitHub repository. Please point this notebook to your local repository of that data. For your convenience this GitHub repository does include the data frozen as of a certain date. The ARIMA model has been hyper parameter tuned, stored on disk and is ready to go. There is a forecasting module that can be ran at any time. It remains to be seen whether or not its predictions will be accurate.

All of the code and data required to run this notebook is in included in the GitHub repository.

Comments

There are a large number of functions in this notebook. Usually I would put those in a helper.py file. However, the specific instructions were to have all code in a single Jupyter Notebook.

Licensing, Authors, Acknowledgements

The John Hopkins data is available under the Creative Commons License. For the particulars of that license please go to https://data.humdata.org/event/covid-19 and search for Creative Commons.

The ARIMA model code was largely taken from this book which I bought "Time Series Forecasting With Python" By Jason Brownlee of Machine Learning Mastery. https://machinelearningmastery.com/make-sample-forecasts-arima-python/ The code was updated to reflect the unique requirements of this covid analysis. Jason provides code with this book that he explicitly tells you to use and make your own.

My thanks to all of the contributors of the data to this project.

About

DSND First Project

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published