Red and White Wine Quality Scores with Physicochemical Properties, 2009

A 2009 dataset combining red and white wine samples for regression modeling of expert-rated quality scores based on physicochemical properties. The data was originally published by Cortez et al. and curated by the TabArena team for a study on independent and identically distributed tabular data. It contains an additional column indicating wine type and is noted to have a high number of duplicate entries.

Use Cases

• Predicting wine quality scores based on physicochemical properties mentioned in the description
• Benchmarking regression algorithms on a combined dataset of red and white wines
• Analyzing the relationship between chemical composition and sensory evaluation outcomes
• Studying the impact of data duplicates on model performance in tabular ML

Strengths

• Licensed under CC BY 4.0 for open use and redistribution
• Based on a peer-reviewed 2009 study with a cited reference
• Combines two original datasets (red and white wine) into a single resource

Limitations

• Column-level documentation is absent; field semantics must be inferred after download
• Row count and file size are unknown, which may limit suitability assessment
• Description metadata is limited; actual data quality requires manual inspection after download

Provenance

Source

Cortez, Paulo, et al. 'Modeling wine preferences by data mining from physicochemical properties.' Decision support systems 47.4 (2009): III.

Collection Method

Curated from the original data source by the TabArena team, combining red and white wine datasets.

Time Range

2009

Freshness

Dataset Year is 2009; last update date is unknown.

Geography

null