TagRefinery: A Visual Tool for Tag Wrangling
summary
TagRefinery is an interactive visual app that cleans open tag spaces using a guided and advanced modes workflow, slider controls for popularity and similarity thresholds, and list-based inspection rather than word clouds. On a 250,000-unique-tag, 5.4M song‑tag pair dataset it reduces to 358 tags from 250,000 and from 5.4M to 1.7M usable pairs, then salvages to 2.3M (≈31% increase). Usability tests (6 novices) and 3 case studies show novices get stepwise help while experts keep granular control; it successfully preprocessed data for a tag‑based recommender.
AI-written summary · about
chapters (4)
related talks
22mEvaluation and Refinement of Clustered Search Results with the Crowd
34mSearching for My Next Chart
08mEmerging Trends in Music
19mPopping the Filter Bubble
35mDiversity and Data: Cases in the Music Industry
08mComparing Apples and Oranges: Human and Computer Clustered Affinity Diagrams Under the Microscope
37mBuilding a Scientific Taxonomy at Scale
29mFrom Keyword Chaos to Clarity: Conquering a Complex Systematic Review with Python