Project 02 · Machine Learning · Streamlit

Access Wildflix

A movie recommendation system based on a KNN model, enriched with data from the OMDB API. Here are the credentials to try the app directly.

User

Password utilisateurMDP

Administrator

Password rootMDP

Test account: shared credentials for the demo, no personal data required.

Wildflix Launch the app

The project

What is it about?

Starting from a movie dataset, the project retrieves enriched information through the OMDB API (genre, cast, director, IMDB rating, poster), then trains a KNN model to suggest five movies similar to the one selected by the user.

Python Scikit-learn (KNN) Pandas API OMDB Streamlit Hugging Face

How it works

Under the hood

Behind the scenes

Data preparation

Dataset import and ID creation from IMDB links
Enrichment through the OMDB API in batches of 1,000 movies
Cleaning, movie filtering, and encoding of genres, actors and directors
KNN model training and export

User side

The Streamlit app

Credential-based authentication (streamlit-authenticator)
Movie selection from a menu (title and year)
Display of the poster, translated plot summary, cast and rating
Recommendation of five similar movies, sorted by IMDB rating

Technical details (repo structure, local setup)

File	Role
`preparation_data_projet2.py`	Data preparation and ML model training
`projet2_V2.py`	Streamlit recommendation app
`bandeau.py`	Wildflix banner display module
`requirements.txt`	Required libraries
`data_patrickV2.csv`	Source dataset
`df_affichage.csv`	Lightweight dataset generated for display

To run the project locally:

Install dependencies with pip install -r requirements.txt
Add an OMDB API key in preparation_data_projet2.py
Launch the app with streamlit run projet2_V2.py

The trained model and the feature matrix are hosted on Hugging Face and loaded automatically by the app.

Team

Built by

Cédric C. Claire G. Elisa G. Eloïse G.