Skip to content

README.md

Latest commit

 

History

History
110 lines (93 loc) · 4.06 KB

README.md

File metadata and controls

110 lines (93 loc) · 4.06 KB

🌱 Climate‑Plan AI
Street‑Level Sensors → Real‑Time Alerts → Personal & Enterprise Climate Decisions

> Prototype status: Front‑end demo running on synthetic + live API data.
> Next milestone: Plug in hardware sensor mesh and launch the AI Planner beta.

⚡ Why We’re Building This

Traditional satellite / city‑station data is too coarse to spot a fire spark, toxic plume, or heat‑stroke hotspot on your block.
Our solution plants low‑cost sensor nodes every 100–250 m, streams the data, and lets AI tell each user—or an entire enterprise—exactly what to do next.

✨ Core Features

Module What it does
Granular Sensor Mesh Virtual for now; future ESP32‑based nodes (temp, humidity, CO₂/PM, flame & gas sensors) every 0.1‑0.25 km.
Early Danger Detection Edge logic flags sudden temperature spikes / gas signatures → instant “Fire / Hazard Possible” alert.
Interactive Map (index.html) Plots every node; markers auto‑tint green/yellow/red. Click to view live metrics + last AI recommendation.
Ops Dashboard (dashboard.html) Buckets all sites into Safe, Moderate, Danger lanes; surfaces impact snippets (schools affected, farm zones, worker clusters).
AI Personalized Planner (road‑map) Advises: “Shift outdoor PE to 7 AM”, “Irrigate Zone A at 4 PM”, “Evacuate aisle 3—possible fire”.
B2B / School API Planned REST+WebSocket layer so partners can ingest insights, embed widgets, and automate ESG reports.

🏗️ Architecture (Prototype → Full Scale)

Sensor Node (ESP32)  ← future  ──MQTT/HTTPS──▶  Ingest API (FastAPI) ─┐
  • Temp / Humidity                                          │ WebSocket / REST
  • CO₂ / PM₂.₅                                              ▼
  • Flame / Gas     (Edge anomaly)                   Front‑End SPA (this repo)


Today the front‑end consumes **Open‑Meteo** + synthetic feeds; swap in real POSTs later.

---

## Quick Start (Prototype)

```bash
git clone https://github.com/YOUR‑ORG/climate-plan-ai.git
cd climate-plan-ai
python -m http.server 8000      # or VS Code Live Server

# Map view
open http://localhost:8000/index.html
# Dashboard
open http://localhost:8000/dashboard.html

Synthetic readings live in dashboard.js → generateFakeReading().

Config Cheat‑Sheet

Tweak Where
Node coordinates monitoredLocations in dashboard.js
Alert thresholds getStatusLabel() in JS files
AI chat canned replies aiResponses in main.js
Live‑Server port .vscode/settings.json

Future Roadmap

  1. Hardware Alpha (10 nodes) – ESP32 + LoRa/4G fallback, solar‑powered.

  2. AI Planner v1 – LSTM / Prophet ensemble for 3‑day, block‑level risk forecasts.

  3. School Pack – Dashboards + push alerts for principals and PE teachers.

  4. Enterprise Tier – SLA dashboards, ESG exports, webhook automations.

  5. Edge‑ML Firmware – Tiny anomaly models on‑device to flag fire/gas in <1 s.

Tech Stack

  • Vanilla HTML / CSS / JS + Leaflet (fast, framework‑free).

  • Open‑Meteo + Nominatim for zero‑key weather / geocoding.

  • (Planned) FastAPI + TimescaleDB for ingest + analytics.

  • (Planned) TinyML on ESP32 for local anomaly detection.

License

MIT — Fork, remix, and deploy anywhere.

Built in 24 hrs for "Let’s Build AI for Climate Resilience(Hackathon)" by Team SkyMinds — “Think globally, sense locally, act instantly.”