SourceTrac: Tracing Data Sources within Spreadsheets

Analyzing data from multiple sources is a common task in scientific research. In particular, spreadsheet data is often aggregated from a variety of sources to identify patterns and synthesize reports. Yet, techniques are lacking for automatically capturing the provenance of such data within spreadsheet environments like Excel. We present a novel approach for fine-grained tracing of tabular data that may have been obtained from files, databases, or the Web. Our approach provides relevant provenance information at both the micro-level (per cell) and the macro-level (per sheet). Initial results suggest that our approach is scalable and beneficial to data analysts.