A static demo site hosting dozens of interactive D3.js visualizations of UFO sighting data. Originally developed by students in USC Data Science courses (primarily DSCI 550 / CSCI 599 Spring 2018) under the direction of Dr. Chris Mattmann as part of the USC Information Retrieval and Data Science (IRDS) Group.
The site demonstrates data collection, enrichment (joining public datasets, OCR on British MoD files, deep learning for image captioning/object recognition via Apache Tika + TensorFlow), similarity analysis (Tika-Similarity), and rich client-side visualization — all made fully self-contained for easy deployment and local viewing.
- http://www.usc.edu/ufodata
- http://irds.usc.edu/ufo.usc.edu/
- GitHub Pages mirror (this repo)
# From the repo root
python3 -m http.server
# Then open http://localhost:8000/
# or http://localhost:8000/html/d3-examples.html for the full galleryNo build step, no external dependencies, no servers required. All data is vendored or synthetic where original backends (Solr/ES) were used.
index.html— Main landing page (About, Insights, Partners, Search/Explore sections)html/d3-examples.html— Central gallery listing all team visualizations with thumbnails and descriptionshtml/team_*.html+teams/team_*/— Individual team hubs (consistent top navbar + cards linking to visualizations)CS599_group_8_final_assignment/— Complete Team 8 (Spring 2018 DATAVIS) submission with its own dashboard-style galleryhtml/,js/,css/,images/,Data/,data_files/— Shared assets, D3 libraries (vendored), and datavisualizations/,d3-visualizations/— Additional standalone examples
Many teams originally relied on dynamic backends. These have been converted to static/synthetic data (JSON/CSV) so everything works reliably from a plain HTTP server.
- Data Enrichment (HW1/HW2 style): Joining UFO sightings with airports, population, weather, etc. + OCR on British UFO files + image analysis with Tika + Inception models.
- Similarity & Clustering: Tika-Similarity (Jaccard, cosine, edit distance), dendrograms, circle packing.
- Interactive D3 Visualizations: Maps (choropleth, dot), timelines, word clouds, gauges, bubbles, circos, radar, heatmaps, image similarity search (MEMEX ImageSpace/ImageCat), Solr-backed dynamic views (now static), and more.
- Consistent Branding: USC/IRDS top navbar on all team pages (Home →
/index.html#page-top, About, Insights, Search, Partners). Iframe loading pattern used on complex hubs so the outer nav is never lost. - Self-Contained Demo: Absolute paths, vendored d3.v3/v4 + topojson, no CDNs, works offline after
python -m http.server.
- Team 2: World map with year filtering + temporal/sci-fi/meteorite viz
- Team 3: US/UK maps (airports + meteorites), scatter, dendro, shapes, word clouds, etc.
- Team 4 (2018 & 2021): Population bars, military, shapes, density maps, bubble maps, time bars, word clouds, sunburst
- Team 5: State maps, time series, gauges, shape donuts, environmental/literacy correlations
- Team 6: Choropleths, shape pies/bars, airport distances, heatmaps, word clouds, state bubbles, climate scatters, elevation dashboard, dynamic bullets
- Team 7: Smog gauges, duration, shape pies/connections, aerial/events, airport distances, word clouds, US maps
- Team 8: Pop culture influence, severe weather, OCR retention (British files), NER, word clouds, sighting date circos, airbase radar, population, Tika vision bubbles (full dashboard gallery)
- Team 10: Word clouds (objects/descriptions/captions), airport/meteor correlations, date diffs, state maps, movie impacts, yearly trends, dynamic Solr
- Team 11: Chords, heatmaps, word trends, 3D shapes, world/US maps, population/marijuana/airport state correlations, image classification
- Team 12: Global distributions, air pollution, airports, population, cancer incidences, shapes/states, top cities, word clouds, world map
(Teams 1, 9, and 13 did not submit complete work and are omitted from the gallery.)
- Dr. Chris Mattmann — http://mattmann.ai
- Student teams from DSCI 550 / CSCI 599 (Spring 2018 and subsequent semesters)
- USC Information Retrieval and Data Science Group (IRDS) — http://irds.usc.edu/
- Tools & libraries: Apache Tika, Tesseract, TensorFlow/Inception (via Tika dockers), D3.js, MEMEX ImageCat + ImageSpace, Solr/Elasticsearch (originally), FLANN, SMQTK, etc.
- Original data sources: UFO-Awesome corpus (example: https://www.kaggle.com/datasets/camnugent/ufo-sightings-around-the-world), UFO Stalker (https://ufostalker.com/), British MoD UFO files
Apache License 2.0 (see LICENSE)
This repository aggregates and maintains student work from multiple semesters. All visualizations have been made self-contained, paths normalized for GitHub Pages + local http.server, consistent navigation added, and non-functional dynamic features replaced with synthetic data so the entire demo "just works."
Pull requests and issues welcome for fixes, new visualizations, or documentation improvements.
Maintained by the USC Data Science Group. Questions? Contact the IRDS team.