Awesome Open Source
Awesome Open Source

Data Visualization for Formula 1 Races

A full-stack dockerized web application to visualize Formula 1 race statistics from 2016 to present, with a Python Flask server and a React front-end with d3.js as data visualization tool.

Hosted at:

I will no longer be updating this repository and the latest code changes will be commited to the above repositories instead.


Data Source

  • Thanks to the Ergast Developer API (, which provides data for the Formula 1 series and is updated after the conclusion of each race.

How to automate the refresh/update of data visualization dashboard?

  • This requires automating the data collection process. To do this, I created a task scheduler within the app powered by Celery to fetch data from Ergast's APIs periodically. Next, I created Python scripts to perform data transformation. The processed data is then saved to a Postgresql database which is hosted on AWS.
  • Celery schedules data to be extracted from Ergast API every Monday morning. If the day before is not a race weekend (Race weekends are spread out from March to November with races occurring on Sundays), nothing gets saved to database and the scheduler retries the following week.
  • The processed data is then retreived from database for implementing APIs.

How to connect Flask and React?

  • I used Flask to create REST APIs and have React consume the APIs by making HTTP requests to it.
  • I did not use the create-react-app library , hence I had to create a custom Webpack configuration. Webpack and Babel (transpiler to convert ES6 code into a backwards compatible version of JavaScript) bundles up the React files in a folder separate from the Flask app.

Data Visualization in React using D3(V4)

  • I used these two libraries together to create dynamic data visualization components. Data is retrieved from the APIs created by Flask.

  • React and D3 are both able to control the Document Object Mode (DOM). To separate responsibilities as much as possible, I went by the following approach:

    • React owns the DOM
    • D3 calculates properties

This way, we can leverage React for SVG structure and rendering optimizations and D3 for all its mathematical and visualization functions.


  • The front-end and back-end was each deployed to separate AWS Elastic Beanstalk environments.
  • I attempted to deploy it as a single AWS EB app, but encountered some issues configuring Nginx to frontend my backend services. Furthermore, the larger app size meant i have to upgrade the EC2 instance type to 't2.small', which had to be paid for. Hence, it was a more viable option to deploy to two separate EC2 instances.
  • The Webpack build to generate static assets happens locally before deployment and the generated files are bundled with the deployment package.



Getting Started


This setup is built for deployment with Docker.

1. Clone the repository

cd ~
git clone
cd celery-scheduler

2. Install Docker

3. Build docker images with docker-compose and run it.

Configuration folder architecture:

│   │
│   └───development
│   │   │   dev-variables.env
│   │   │   docker-compose.yml
│   │ 
│   └───production
│       │   .env
│       │   docker-compose-prod.yml
│ Dockerfile
│ Dockerfile-node

To test the app locally, first enter the correct folder.

cd config/docker/development

Then execute the following command:

docker-compose -f docker-compose.yml up -d --build

I have configured Docker such that when the postgres image is built and an instance (container) of it runs, a new database is created along with a postgres user and password. However, the database is currently empty and requires a psql script to load it with some data. The database shuts down when the container stops and is removed.


  • -f: specify docker-compose file name (Not necessary to specify, unless named differently from standard 'docker-compose.yml'
  • up: Builds, (re)creates, starts, and attaches to containers for a service.
  • -d: Detached mode: Starts containers in the background and leaves them running
  • --build: Build images before creating containers.

4) Check logs for successful build and run of docker containers

  docker-compose logs

Please refer to this repo's wiki for screenshots of what you should see from the console.

5) Loading database with data

I am unable to succesfully use an entrypoint script to initialize database with data, hence the workaround will be to manually load data from command line instead.

a. Check the list of running containers

  docker ps -a


b. To run bash command in docker container, enter

  docker exec -i -t <CONTAINER_ID> /bin/bash

In this case, we want to enter the 'development_postgresql' container, so run docker exec -i -t 899f05bf2ec2 /bin/bash (Note: my CONTAINER ID will be different from yours)

c. Run below command to dump 'init.psql' to the database

  psql --host=localhost --port=5432 --username=test_user --password --dbname=f1_flask_db < ../init.psql

You will then be prompted for the password for test_user, which is 'test_pw'

d. Log into the database. Try querying it!


You may now point your browser to http://localhost:3000 to view the frontend, and to http://localhost:5000/api/results to view the APIs

6) Initialize task scheduler with celery

a. Based on the last step, we should still be in 'development_postgresql' container. Exit from postgresql by entering '\q'. Exit from container by entering 'exit'. Next, identify the "development_app" container id and enter it (similar to steps 3b.1 and 3b.2).

b. Trigger the below command:

  celery -A app.tasks worker -B -l info

That's it! For testing purpose, i have set celerybeat to trigger task to data collect every 15 minutes.

7) To stop running of docker containers and remove them

  docker-compose down

For enquiries, you may contact me at [email protected]

Get A Weekly Email With Trending Projects For These Topics
No Spam. Unsubscribe easily at any time.
Javascript (1,512,666
Reactjs (72,763
Docker (33,824
Flask (9,792
Data Visualization (4,374
D3js (1,090
Sports (453
D3 Visualization (194
D3v4 (164
Related Projects