Engineer Insightful Shiny Apps with Behavioral Science
summary
Dashboard failures often come from psychology not code: a 30 day telemetry trace shows 95% open the URL, 88% land on the app, half use only one filter and 20% return after a week. The BID Framework prescribes five stages (interpret, notice, anticipate, structure, validate), maps theories to UI fixes like bslib accordions and F‑pattern layouts, and the bidux R package (CRAN) autogenerates suggestions, bid_report() output, and uses telemetry plus fuzzy dictionary matching (no LLMs); after changes filter usage rises 38% and return rate 75%.
AI-written summary · about
chapters (13)
- 0:00Welcome
- 1:07Psychology to dashboard engineering
- 3:17Human behavior in dashboards
- 4:38Behavioral science principles for UX
- 6:25BID framework
- 7:28Interpret stage: user narrative
- 8:34Notice stage: friction points
- 11:07Structure stage: information flow
- 12:34Validation stage: user empowerment
- 13:36Fixing a broken dashboard with BID
- 15:48Implementing BID with bidux
- 17:13Resources and getting started
- 18:22Q&A on bidux privacy and Quarto
related talks
27mVisual Design - Do not let Data be your cul-de-sac!
35mApplied Psychology: Psychology based UI improvements
41mEffective Dashboard Design Using
03m"A Story about Design and Experience"
49mWeb psychologist explains the secrets of addictive UI design
37mBehind the AI Curtain - Designing for Machine Learning Products
40mButtons hidden in plain sight
50mDVUX: Data Visualization and User Experience