ClimateLens

Engineering Intelligence · Software Development · Applied AI

ClimateLens is a FastAPI service and interactive web dashboard for climate-risk screening and Earth-system trend analysis. It combines a NASA POWER data client, heuristic risk scoring, and statistical trend detection (Mann-Kendall, Theil-Sen) in a single, self-contained application.

Repository: https://github.com/fazeli1977-dot/ClimateLens

Overview

ClimateLens exposes two analysis endpoints and a browser-based dashboard:

• Climate risk analysis (/api/v1/risk-analysis): retrieves daily temperature, humidity, and precipitation for a coordinate and lookback window, then computes a 0-100 risk score with thermal-stress, drought, and volatility components.
• Earth-system trend detection (/api/v1/trend-detective): returns Mann-Kendall significance, Sen's slope per decade, baseline drift, and a scientific verdict for a multi-decadal series.

The application is a prototype/research tool. It is not a validated operational climate-risk service, and its scores and trends should be treated as exploratory indicators rather than authoritative forecasts.

Key capabilities

• NASA POWER daily point data client (T2M, RH2M, PRECTOTCORR) with a local synthetic fallback when the upstream request fails.
• Heat-index approximation and a 0-100 composite risk score (thermal stress, drought, temperature volatility) with Low / Moderate / High / Critical levels.
• Mann-Kendall trend test and Theil-Sen (median) slope estimation, including Kendall's tau, p-value, and decadal slope.
• Interactive dashboard: coordinate selection, risk score and advisory, metric grid, daily time-series chart, and trend chart.
• Automated tests for risk scoring and trend detection.

Project structure

ClimateLens/
├─ backend/
│  ├─ app/
│  │  ├─ main.py
│  │  └─ services/
│  │     ├─ nasa_service.py
│  │     ├─ risk_engine.py
│  │     ├─ risk_service.py
│  │     └─ trend_detector.py
│  └─ tests/
│     └─ test_risk.py
├─ frontend/
│  └─ public/
│     └─ index.html
├─ tests/
│  └─ test_trend_detector.py
├─ ClimateLens_Presentation.pptx
├─ NASA_Submission_Details.txt
├─ pytest.ini
├─ .gitignore
└─ README.md

Tech stack (verified)

• Python 3.12
• FastAPI, httpx, Uvicorn
• Vanilla HTML/CSS/JavaScript dashboard (no build step)
• pytest (development/testing)

No dependency manifest (requirements.txt or pyproject.toml) is present in the repository.

Quick start

1. Clone

git clone https://github.com/fazeli1977-dot/ClimateLens.git
cd ClimateLens

2. Virtual environment

python -m venv .venv

Windows:

.venv\Scripts\activate

macOS/Linux:

source .venv/bin/activate

3. Install dependencies

pip install fastapi httpx uvicorn

For tests:

pip install pytest

4. Run

python -m uvicorn backend.app.main:app --reload

Open http://127.0.0.1:8000/ for the dashboard and http://127.0.0.1:8000/docs for the API documentation.

API

GET /api/v1/risk-analysis

Query parameters:

• lat (float, -90 to 90, required)
• lon (float, -180 to 180, required)
• days (int, 7 to 90, default 30)

Returns location, elevation, risk assessment (score, level, metrics, breakdown), and daily climate history.

GET /api/v1/trend-detective

Query parameters:

• lat (float, default 35.6892)
• lon (float, default 51.3890)
• variable (string, default "T2M"; "T2M" selects the temperature baseline, otherwise precipitation)
• start_year (int, default 2000)
• end_year (int, default 2024)

Returns target, timeline, Mann-Kendall statistics, Sen's slope per decade, drift, percentage change, trend classification, scientific verdict, and data points.

Note: this endpoint currently derives its series from a deterministic in-code baseline rather than a live historical dataset.

Data sources

• NASA POWER (Prediction of Worldwide Energy Resources) daily point API: T2M, RH2M, PRECTOTCORR, community RE. Used by /api/v1/risk-analysis.
• When the upstream request fails or returns no data, a local synthetic series is generated from latitude-based heuristics. These values are not measurements.
• The trend-detective endpoint uses a deterministic synthetic series defined in code; it does not query NASA POWER or other providers.
• Submission materials (NASA_Submission_Details.txt) reference additional NASA Earth-observation concepts (GPM, MODIS/VIIRS) and Open-Meteo, but these are not implemented as data integrations in the current codebase.

Users are responsible for complying with upstream provider terms when integrating or redistributing data.

Testing

python -m pytest -q

The suite contains 4 tests covering heat-index calculation, risk evaluation, and Mann-Kendall trend detection (positive trend and stable series).

Scope and limitations

• Risk scores and advisories are heuristic indicators, not validated predictions or operational alerting.
• The NASA POWER fallback and trend-detective series are synthetic/generated, not observations.
• No authentication, rate limiting, or production hardening is implemented; CORS is permissive (*).
• No dependency manifest or container configuration is included.

Contributing

Open an issue to discuss major changes, then submit a pull request with a clear description.

License

No LICENSE file is present in this repository, and no license terms are specified. License information is therefore missing; until a license is added by the owner, default copyright protections apply. Users should contact the repository owner before reuse or redistribution.

Developer

Developed by Alireza Sani as part of an international engineering and software portfolio.

Portfolio: https://github.com/fazeli1977-dot