Crowdsourced Wi-Fi Fingerprinting Database for Indoor Positioning in a University Building

Two Wi-Fi fingerprint databases for training and testing indoor positioning algorithms, collected via 21 different devices in a crowdsourced mode. The data comprises 687 training fingerprints and 3951 test fingerprints, collected in a 4-floor university building in Tampere, Finland, during January to August 2017. The dataset was created by researchers including Elena Simona Lohan from Tampere University and includes benchmark utility software in Matlab and Python.

Use Cases

• Benchmarking indoor positioning algorithms based on Wi-Fi MAC addresses and RSS values.
• Training machine learning models for location estimation using crowdsourced fingerprint data.
• Evaluating the performance of positioning systems across different user devices in a multi-floor environment.
• Studying the variability of Wi-Fi signals in a real-world indoor setting collected over several months.

Strengths

• Contains 687 training and 3951 test fingerprints, providing a clear split for model development and evaluation.
• Data was collected via 21 different devices and different users, suggesting diversity in the signal capture.
• Includes benchmark utility software in Matlab and Python to illustrate various positioning algorithms.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

Tampere University

Collection Method

Crowdsourced collection via 21 different devices and users.

Time Range

Jan-Aug 2017

Geography

A 4-floor university building in Tampere, Finland