Extracting Format Transformation Examples from Manual Data Corrections.

One of the challenges in data analysis is the substantial cost of human involvement. Before any analysis can take place, data from heterogeneous sources must be cleaned, integrated and transformed into a uniform format. These tasks often require both technical skill and domain expertise. Current tools support data scientists by allowing them to define rules to carry out the preparation work, but the problem of the manual work needed to create such rules remains. We propose an approach that observes the updates made by data scientists when manually correcting errors in query results, and uses them to derive format transformation rules that can be applied to this and future query results.