Live COVID-19 tracker with Airflow

Note: The app is no longer live or maintained. To run it locally you may clone the repository, install all requirements in requirements.txt, and do

streamlit run 3.0App.py

• The final live streamlit app looked like in the following screenshot:

• The final state (2020/09/25) of a previous version is still accessible in this webpage.

---

Set of charts automatically updated daily with Apache Airflow. COVID-19 data provided by the Johns Hopkins University Center for Systems Science and Engineering (JHU CSSE). The data can be found in this GitHub data repository.

• Link to the live tracker

• A Medium Story of this project was featured in the Data Science and Programming topic pages, and was published in Towards Data Science

We use the bokeh and plotly visualization libraries.

2.0COVID19_notebook.ipynb

Along the notebook, we

1. Load and clean the data.
2. Show bokeh interactive bar plots for the top countries by confirmed cases, deaths, recoveries and mortality rate.
3. Present the world totals.
4. Compute the daily cases and show bokeh interactive time series plots.
5. We show plotly geographical, interactive maps.

Automated/

Series of scripts used for Airflow:

• covid19_dag.py Airflow DAG that automates the execution of:
  • covid_func.py Reproduces the code in the Jupyter notebook.
  • git_push.py Commit/push of the plots to the GitHub Pages repository.

The tracker was updated daily at 05:00 UTC.