AAmogh·Panhale
All projects
AndroidResearch Data2026

TikTok VASP

Android research platform that captures scrolling behavior in a TikTok-style feed.

Code

Problem

The VASP Lab at the University of Alberta needed to study short-form-video scrolling behavior, but off-the-shelf apps don't expose the granular touch data psychology research requires.

What I built

  • Built a full TikTok-style vertical feed in Kotlin with Jetpack Compose + Media3 ExoPlayer and auto-advancing ViewPager2 playback.
  • Wrote a swipe-analytics service that captures the full touch path and derives velocity, acceleration, jerk, straightness, and smoothness — converting pixels to meters via device DPI for cross-device comparability.
  • Added researcher-configurable sessions (participant ID, video set, duration, randomized input-blocking interruptions) and automatic CSV/JSON export of per-view and per-session data.

Impact

  • Deployed for participant studies in the VASP Lab.
  • Exports structured play-by-play and session datasets ready for analysis.

Stack

KotlinJetpack ComposeMedia3 ExoPlayerMVVMViewPager2