Demo Guide

Demo Guide

Try the tool without setting up your own data.

Live Demo

The live demo is hosted on Hugging Face Spaces:

https://tavoloperuno-accelerometry-viewer-demo.hf.space/

Credentials

Username	Password	Role
demo_admin	demo	Admin — can impersonate users, manage users
demo_user	demo	Annotator — standard annotation workflow

What to Try

1. Log in as demo_user to see the annotator experience:

   • Select a file from the dropdown
   • Navigate using Previous/Next buttons or the minimap
   • Box-select a region and click an activity button
   • Toggle flags (Segment, Scoring, Review)
   • Add a note and click Export

2. Log in as demo_admin to see admin features:

   • Use the Impersonate dropdown to view demo_user's annotations
   • Open the User Admin panel to see user management
   • Review pre-populated annotations showing all activity types and flags

Demo Data

The demo uses real tri-axial accelerometer data from the UCI Accelerometer Gyro Mobile Phone Dataset (CC BY 4.0), resampled to 85 Hz and composed into ~10-minute files. Pre-populated annotations showcase all activity types, flags, and inter-annotator variability.

Running the Demo Locally

Generate demo data

# Real data (downloads ~2 MB from UCI)
python demo/generate_data.py

# Synthetic fallback (sine waves, no download needed)
python demo/generate_data.py --synthetic

# Custom output directory
python demo/generate_data.py /tmp/demo_output

Start the demo server

panel serve demo/app.py \
    --port 7860 \
    --basic-auth demo/credentials.json \
    --cookie-secret $(python -c "import secrets; print(secrets.token_hex(32))") \
    --allow-websocket-origin localhost:7860 \
    --basic-login-template visualize_accelerometry/templates/login.html

Open http://localhost:7860/app and log in with the credentials above.

Docker

docker build -f demo/Dockerfile -t accelerometry-demo .
docker run -p 7860:7860 accelerometry-demo

Open http://localhost:7860/app.