This repository shows some interesting video analytics from YouTube based in United States for the year 2017 and 2018. These analytics are based on the stats file from Kaggle here
The repository has Jupyter Notebook to visualize the stats in browser and also has a FastAPI app to serve some interesting stats using REST endpoints. Both the Jupyter and FastAPI source code is dockerized and could be run through docker compose to avoid running manual steps for installing dependencies, CLIs and servers.
- Python 3.11
- Pandas
- Matplotlib
- Numpy
- FastAPI
- SQLAlchemy
- SQLite
- Jupyter
- Uvicorn
- Docker
Follow these steps to setup and run both FastAPI app and Jupyter notebook
Clone this repo
git clone https://github.com/lalatgithub/youtube-trending-video-analytics.git
Switch to cloned dir and run.
cd youtube-trending-video-analytics
NOTE: If you want to run FastAPI app and Jupyter on a different port, Please update .env accordingly in root dir.
docker compose up
once docker compose runs successfully, you should be able to fetch data from REST endpoints here http://localhost:8000/. You can see that you have a ./analytics.db file in the project root dir.
The analytics are loaded from ./datasets/USVideos.csv during the app startup.
The Jupyter notebook can be access here http://localhost:8888/notebooks/YouTube%20Trending%20Videos.ipynb