European Personal Location Information for PII Masking

A preview of 50 sample entries from the PII-Masking-2M European release focused on Personal Location and Travel Information (PLI). The dataset is produced by AI4Privacy, with the preview published in April 2026. Full access to the complete 200,000-entry dataset requires contacting the authors.

Use Cases

• Training models to identify and mask location-related PII entities like addresses or travel routes using the annotated text structure.
• Benchmarking the performance of text anonymization tools on European language data, as indicated by the language distribution metadata.
• Analyzing patterns in European location and travel information for privacy risk assessment, leveraging the geographic coverage metadata.
• Developing synthetic data generation techniques by studying the label distribution of redacted PII values within the sample entries.

Strengths

• Contains 50 sample entries providing a concrete preview of data structure and annotation format.
• Part of a larger 2M-entry collection, suggesting a substantial underlying data source.
• Explicitly documents European coverage and language distribution for geographic and linguistic context.

Limitations

• Sample size is limited to 50 rows, insufficient for model training or statistical analysis.
• Source text and actual PII values are redacted in the preview, preventing direct use for model development.
• Key details like column names, total row count for the full release, and file formats are unknown.

Provenance

Source

AI4Privacy's PII-Masking-2M European release.

Collection Method

Presumably collected and annotated for PII masking research; specific method is null.

Freshness

Last updated in April 2026.

Geography

European coverage.