This project implements a data-driven decision-making process for laptop selection using a weighted-scoring method.
It combines real laptop specs with benchmark data and converts them into a single comparable score.
- Data Collection
- Specs are collected manually from Public.gr (Greek retailer) during Aug 2025.
- Only laptops that were in stock and meet minimum requirements are included.
- Feature Categories
- Quantitative: CPU benchmark, RAM size, storage size, battery life, weight, price, connectivity (USB, HDMI etc).
- Categorical: Storage type (NVMe, SSD, UFS), RAM type (DDR4, DDR5, LPDDR5, LPDDR5X), Display panel (IPS, OLED, etc.), Display resolution (HD, FHD, QHD, 4K), Wireless technology and type, Webcam resolotion (720p -- 2160p).
- Normalization
- Min-Max normalization is used to scale all features between 0 and 1.
- Cost-type features (like weight, price) are reversed so that lower = better.
- Weighted Score
- Each feature is assigned a weight (chosen features sum to 1.0).
- Final score is calculated as a weighted sum of all normalized features.
- Ranking
- Laptops are sorted by their final score.
- The top-ranked laptops are the best balance of performance, portability, and value.
The project includes a simple bar chart comparing the Top-5 laptops by score:
- Collect raw laptop data into
laptop_data.csv. - Run the pipeline:
df = pd.read_csv('laptop_data.csv')
df = data_transformation(df)
df = get_scores(df)
df = rank_laptops(df)- Review the ranked list and generate a chart.
- Benchmarks are taken from PassMark (CPU).
- Features used can be adjusted depending on the users preferences.
- Weights can be adjusted depending on user needs (e.g. more focus on battery life or portability).
- This is a real world project and helps individuals to make a data-driven laptop choices yet it can be applied to any choices.
The project's image was generated by @DALL·E 3
🤗 Thank you for visiting this project, and I hope it will help you make emotionless decisions 😊