---
license: mit
pretty_name: chess
tags:
- chess
- recommendation system
size_categories:
- n<1K
---

# Chess Opening Recommendation Dataset

This is a small dataset representing a user-opening interaction matrix (with 20 users and 4418 different opening variations), designed for training a recommender system to suggest chess openings. 
Each row corresponds to a user, and each column represents a specific chess opening or variation.
The dataset was created using information gathered using the Chess.com API, therefore all users in the dataset are from Chess.com.

## Dataset Structure
- Rows: Users
- Columns: Chess openings and variations (according to ECO)
- Values: Lists containing two percentages:
  - Frequency: How often the user plays the opening as a percentage.
  - Win Rate: The success rate of the user with that opening as a percentage.


## Usage

This dataset can be used to develop and evaluate chess opening recommendation models

Enjoy and happy modeling!