Nomad

A Next.js research-mapping app that ranks papers, researchers, subtopics, citation signals, and source-linked project directions from an engineering query.

Quick start · Evaluation · Data and visuals · Configuration

Overview

Nomad turns a broad literature query into a smaller starting map: papers to read, recent work to watch, researchers, topic clusters, a reading roadmap, and project directions linked back to source records. The classic map is at /; the v2 explorer at /explore adds a knowledge graph, mind map, trend radar, and quality report.

flowchart LR
    Q[Topic, field, goal, years] --> V[Query variants]
    V --> O[OpenAlex works and authors]
    O --> N[Normalize, deduplicate, and filter]
    N --> S[Rank and cluster]
    E[Optional OpenAI embeddings] --> S
    S --> M[Research map and quality report]
    S --> X[Graph, roadmap, mind map, and trends]

Loading

OpenAlex is the primary data source. Nomad works without an OpenAI key by using deterministic keyword, metadata, and citation signals. Citation-based ranks are reading-order signals, not predictions of research value or future impact.

Quick start

npm ci
npm run dev

Open http://localhost:3000 or http://localhost:3000/explore, then enter a topic such as machine learning for HVAC CFD. Queries that name a method, system, application, or measurable outcome usually map more cleanly than a broad field name.

Run the local quality gates with:

npm run lint
npm test
npm run build

The test suite uses frozen OpenAlex fixtures where available; it does not need network access or an OpenAI key. Current fixture coverage and gaps are listed in docs/evaluation-results.md.

Configuration

Variable	Required	Purpose
OPENALEX_MAILTO	No	Adds a contact address to OpenAlex requests
OPENAI_API_KEY	No	Enables semantic relevance scoring

Copy .env.example to .env.local. Do not expose either value through a NEXT_PUBLIC_ variable.

Documentation

Document	Purpose
docs/evaluation-results.md	Current automated evaluation coverage, results, and missing evidence
docs/data-and-figures.md	OpenAlex lineage, ranking path, and runtime visualization production
docs/figure-manifest.json	Machine-readable map for diagrams and runtime visuals
PAPER_SELECTION_FRAMEWORK.md	Paper-selection logic
PAPER_RANKING_DEV_PLAN.md	Ranking implementation plan
RELEVANCE_JUDGE_PLAN.md	Optional relevance-judge plan
ROADMAP.md	Release boundary and deferred work

Repository map

app/page.tsx              classic research-map interface
app/explore/              v2 explorer and visualization components
app/api/research-map/     classic and v2 API routes
lib/openalex.ts           OpenAlex retrieval, caching, and normalization
lib/scoring.ts            deterministic ranking and project synthesis
lib/derive/               graph, mind map, roadmap, trends, and explanations
lib/quality/              relevance, subcluster, trend, and regression checks
test-fixtures/openalex/   frozen OpenAlex captures for topic-level evaluation
docs/                     evaluation and visualization lineage

See CONTRIBUTING.md for test, fixture, and UI requirements. Nomad is available under the MIT License.