Algerian Laptop Market
Data mining project analyzing the Algerian laptop market: web scraping, EDA, price prediction with regression models, association rule mining, and a deployed Django price predictor.
Project Overview
An end-to-end data mining project on the Algerian laptop market, built from 16,283 Ouedkniss listings. It covers web scraping of specifications and prices, preprocessing with CPU and GPU benchmark enrichment, exploratory analysis with geographic and temporal visualizations, and several regression models for price prediction.
The pipeline runs from preprocessing (cleaning, feature engineering, segmentation) through exploratory analysis, a comparison of regression models (Linear, Ridge, Lasso, Random Forest, Gradient Boosting, SVR), and association rule mining with the Apriori algorithm.
A Django web application was built and deployed to give real-time price predictions from a laptop's specifications. The Gradient Boosting Regressor performed best, reaching R² = 0.90, MAPE = 15.2%, and MAE = 17,678 DZD on the held-out test set, which makes it a practical pricing tool for the Algerian market.
Key Features
- Web scraping for laptop data and PassMark CPU/GPU benchmarks
- Preprocessing pipeline with feature engineering and enrichment
- Exploratory data analysis with GeoPandas and statistical visualizations
- Price prediction with multiple regression models (R² ≈ 0.90)
- Association rule mining (Apriori) for feature patterns
- Deployed Django web app for real-time price prediction
Technologies Used
Project Details
Client
Academic project for the Data Mining course at ENSIA
Timeline
2025–2026
Role
Data Scientist & Full-Stack Developer
More Projects
© 2026 Yassir CHERDOUH. All rights reserved.