Answer: (i) Geographic-based data

Spatial data mining is the process of discovering interesting, useful patterns from large spatial (geographic) databases — i.e., data that references objects by their location in space, such as maps, satellite imagery, GPS traces, and location-tagged records.

Examples of spatial data:

• Satellite remote-sensing images (land cover, weather patterns),
• Mapping data from Geographic Information Systems (GIS),
• Location data from mobile phones/sensors,
• Earthquake or disease-spread records with co-ordinates.

Techniques include spatial clustering (e.g., finding densely populated regions), spatial outlier detection (e.g., detecting anomalies in pollution monitors), and spatial association rules (e.g., "regions near a river usually have fertile soil").

Why the others are wrong:

• (ii) Text documents — that is text mining.
• (iii) Financial data only — that is financial data mining; spatial mining is not limited to financial data.
• (iv) Audio signals — that is multimedia mining.

Hence spatial data mining deals with geographic-based data.