Abstract
We present CoWrangler, a real-time data wrangling recommender system, which can recommend the next-best data wrangling operations along with the corresponding human-readable and efficient code snippets to expedite data exploration and wrangling efforts. A key feature of CoWrangler is that it provides explanations for the generated suggestions in the form of data insights, allowing the user to place confidence in the system. Under the hood, CoWrangler relies on intelligent generation of candidate suggestions using program synthesis techniques and ranking of a set of suggestions based on the notion of data quality improvement. We demonstrate how CoWrangler provides a human-in-the-loop data wrangling experience, and helps users make informed data pre-processing decisions, while saving their time and effort.
Original language | English |
---|---|
Title of host publication | SIGMOD '23: Companion of the 2023 International Conference on Management of Data |
Pages | 147-150 |
Number of pages | 4 |
DOIs | |
Publication status | Published - 5 Jun 2023 |
Externally published | Yes |