UFC Datalab: 6,000 MMA Fights and 144 Performance Metrics

Maintained by komaksym and updated in November 2025, this repository provides UFC fight statistics and event data gathered through web scraping and OCR. The project focuses on the extraction and analysis of mixed martial arts match records for data science applications.

Use Cases

• Exploratory data analysis of UFC fight outcomes using historical match statistics
• Training predictive models for winner classification based on OCR-extracted event data
• Benchmarking the accuracy of sports-specific parsing and scraping scripts

Strengths

• MIT licensed for open redistribution and modification
• Updated in late 2025 ensuring recent event coverage
• Utilizes OCR to capture data from non-tabular visual sources

Limitations

• Unknown record count and specific column schema
• Potential for character recognition errors inherent in OCR-extracted text
• Lack of sample data for immediate validation of data density

Provenance

Source

UFC (Ultimate Fighting Championship) event records and broadcasts

Collection Method

Scraped and parsed using OCR and web scraping tools

Freshness

Updated as of November 18, 2025.

Geography

Global