Wordcloud

A browser-based visualization project that turns curated French lexical data into a navigable graph, combining deterministic layout, custom canvas rendering and learner-facing exploration.

Product problem

Vocabulary lists are easy to collect but hard to navigate. Learners often need to understand not only what a word means, but how it connects to families, contrasts, clusters and nearby concepts in a larger structure.

Key engineering decisions

The project treats vocabulary as graph data rather than flat content. Curated lexical sources are processed into a stable graph model, grouped through Louvain community detection, positioned through ForceAtlas2 and exported into a deterministic browser runtime.

Interaction model

The browser interface uses custom canvas rendering for the global map and focused word exploration. Learners can move between overview and single-word focus, then keep personal review, draft-card and learning state locally in the browser without requiring an account system.

Technical scope

Graph data modelling, layout computation, browser-side canvas interaction, lexical-source curation, deterministic export, local study-state design and static deployment packaging.

Curated lexical data
  ↓
Graph build → Louvain communities → ForceAtlas2 layout
  ↓
Deterministic export → Canvas runtime → learner exploration