Retail Purchase Data from a Turkish Supermarket with Macroeconomic Indicators

65,895 rows of retail transaction data from a supermarket operating in Türkiye, integrated with monthly macroeconomic indicators. The dataset contains 106 anonymized variables, prepared for machine learning and time-series analysis. It was authored by Osman Nuri TEKİNAY and last updated on 2026-05-08.

Use Cases

• Forecasting retail sales based on integrated macroeconomic indicators
• Training artificial neural networks (ANN) on anonymized transaction patterns
• Conducting causality analysis between interest rates and consumer purchases
• Performing time-series analysis using the provided time index and trigonometric date features

Strengths

• 65,895 rows provide a substantial sample for analysis
• 106 variables offer a wide range of anonymized features for modeling
• Integration of macroeconomic indicators with transaction data enables multi-factor analysis
• Data is preprocessed with logarithms applied to most variables for machine learning readiness

Limitations

• Column-level documentation is absent; field semantics must be inferred after download
• Data may reflect geographic and temporal bias inherent to a single supermarket in Türkiye
• Row count for the primary transaction table is unknown, which may limit suitability assessment

Provenance

Source

Anonymized data from a supermarket operating in Türkiye.

Collection Method

Data was obtained from retail operations and integrated with external macroeconomic indicators.

Freshness

Last updated 2026-05-08 08:44:05; freshness should be verified

Geography

Türkiye