Mag-Array Explorer

Browser-based interactive explorer for the 30-magnetometer-array indoor localization datasets released alongside:

Position and Orientation Estimation Uncertainty Using Magnetometer Arrays for Indoor Localization Thomas Edridge and Manon Kok, IEEE Journal of Indoor and Seamless Positioning and Navigation, vol. 3, pp. 152–164, 2025. doi: 10.1109/JISPIN.2025.3567258

🔗 Live demo: https://gisbi-kim.github.io/mag-array-explorer/ 📄 Paper github: https://github.com/Tedridge/magnetometerArrayOdometryAnalysis 🗂 Original data: https://hendeby.gitlab-pages.liu.se/research/magdata/

---

What it does

Helps a researcher decide whether the dataset fits their needs in under a minute:

• 3D pose trajectory with optional RGB orientation frames (density / length / width controllable)
• Per-sensor magnetic-field heatmap (30 sensors × time)
• Mean |B|, IMU |acc|/|gyr|, baro height — time-aligned
• Per-sequence metadata: kind, duration, GT availability, anomaly richness (σ |B|), path length, mean/max speed
• Gallery grouped by dataset release — click any thumbnail to switch
• Linked brushing: zooming the heatmap highlights the matching segment on the 3D trajectory in yellow
• |B| colormap range presets matching paper sections IV-B / IV-C for fair cross-sequence comparison

40 sequences across 3 dataset releases:

Tag	Release	Count	Used in paper
TINY	Tiny squares — 0.4 × 0.4 m	9	Section IV-A & IV-B
SM	Small / medium squares — 4 × 4 m	2	Section IV-C (LP-1, NP-1 from MAINS [Huang 2024])
S25	Summer 2025 — complex environments	29	Follow-up data (corridors / snake / spiral / 8-pattern …)

Stack

• Pure HTML + CSS + vanilla JS — no build step, no framework
• Plotly.js for 3D scatter, heatmap, time-series
• Data exported as JSON from a local SQLite DB (see build_static.py)
• Hosted on GitHub Pages

Files

index.html           UI structure
style.css            light theme, Inter font
app.js               state, rendering, linked brushing
build_static.py      SQLite → JSON exporter (one-time)
data/                generated artifacts (~10 MB total)
  index.json         sequence index + gallery thumbnails
  <seq>.json         per-sequence figures payload (~280 KB, lazy-loaded)

Local preview

fetch() is blocked on file://, so use a static server:

python -m http.server 8000
# open http://localhost:8000

Rebuilding the data

If you have the original .mat files, the process is:

1. Download the three dataset releases from https://hendeby.gitlab-pages.liu.se/research/magdata/ (magArrayData.zip, smallSquares.zip, the Summer 2025 set)
2. Combine them into one SQLite DB (the converter script lives in the parent project — not included here)
3. Run python build_static.py to refresh data/

The DB is intentionally not checked in (~360 MB).

Citation

If this explorer helps your work, please cite the underlying paper:

@article{edridge2025magarray,
  author  = {Edridge, Thomas and Kok, Manon},
  title   = {Position and Orientation Estimation Uncertainty Using Magnetometer Arrays for Indoor Localization},
  journal = {IEEE Journal of Indoor and Seamless Positioning and Navigation},
  volume  = {3},
  pages   = {152--164},
  year    = {2025},
  doi     = {10.1109/JISPIN.2025.3567258},
}

Acknowledgements

Data collected by Thomas Edridge & Manon Kok (TU Delft) in collaboration with Gustaf Hendeby (Linköping), Isaac Skog & Chuan Huang (KTH).