Telescope — Zoom into your graph

A visual-analytics dashboard for the first exploration of attributed knowledge graphs. Load a graph and get its main metrics — degree distribution, centralities, components, type mixing, timelines — computed, visualised, and interactively explorable. No notebook, no code.

Developed by students of the Course on Visual Analytics (University of Pisa) for the VAST Challenge 2025 – Design Challenge. The challenge asks for a visual-analytics design that helps non-expert users explore knowledge graphs: discover new relationships, identify anomalies, and infer missing information.

Challenge page: https://vast-challenge.github.io/2025/DC.html

Quick start

./dev.sh            # backend (:8000) + frontend (:5173); Ctrl+C stops both
./dev.sh --dev      # also install test deps and run the API suite first
./dev.sh --log      # stream full pip/npm output

The script creates api/venv, installs Python and Node dependencies on first run, then starts uvicorn and Vite. It needs a Linux shell with bash, python3, and Node (^20.19.0 || >=22.12.0). To start backend or frontend by hand, see the README in each folder.

Repository layout

Path	What's inside
frontend/	Vue 3 + D3 single-page app: the dashboard, its panels, stores, and composables.
api/	FastAPI backend: graph loading, indices, and per-panel analysis endpoints.
docs/	Design documentation: what Telescope is, the cross-panel contract, meeting history.
data/	Dataset download helper. The datasets themselves are not committed.

Documentation

Topic	Where
Design rationale, audiences, scope	docs/README.md
Cross-panel interaction contract	docs/contract.md
Backend architecture, endpoints, setup	api/README.md
Frontend architecture, conventions, setup	frontend/README.md
Panel catalogue and authoring guide	frontend/src/panels/README.md
Meeting notes and early wireframes	docs/history/

Stack

Area	Tech
Frontend	Vue 3 · Vite · Pinia · Vue Router · Tailwind CSS v4 · D3 · Lucide · @vueform/slider
Backend	FastAPI · NetworkX · NetworKit (centralities) · powerlaw (distribution fits)

Team

Developed collaboratively by students of the Course on Visual Analytics. Specific contributions, from the git history:

• Salvo Rinzivillo — project scaffolding (Vue 3 + Vite setup, Tailwind integration, first layout components), the original upstream API and its test suite, the initial README, meeting 1 notes, and the data download script.
• Giulia Fabiani (@g-fabiani4) — dashboard rework into nested router views (Graph/Guide), responsive layout, accessibility pass, Graph-view panels and per-view panel filtering, dataset-view loading skeletons, meeting 4 notes and wireframes.
• Francesco Secoli (@sclfnc) — the Telescope reference implementation: the modular FastAPI backend (indices, analysis endpoints, centrality pipeline), the D3 panels, the filter/selection bitmap model and cross-panel contract, the design system, and the design documentation.

References

• VAST Challenge 2025 – Design Challenge: https://vast-challenge.github.io/2025/DC.html
• IEEE VIS / VAST community resources on visual analytics, graph visualization, and knowledge graphs