- What is This Project?
- Why is This Important?
- Key Features
- Screenshots
- Tech Stack
- Quick Start
- Project Structure
- Data Flow & Workflow
- SDG Alignment
- API Reference
- Frequently Asked Questions
- Contributing
- License
- Acknowledgments
Kerala EV Charging Infrastructure Optimizer is an interactive geospatial decision-support platform that helps identify optimal locations for electric vehicle (EV) charging stations across Kerala, India.
The platform combines multi-layer spatial analysis, real-time navigation, and data visualization to serve:
| User Type | Use Case |
|---|---|
| 🏛️ Government Planners | Strategic infrastructure investment decisions |
| ⚡ Charge Point Operators | Site selection for new charging stations |
| 🏢 Real Estate Developers | EV-readiness assessment for new projects |
| 🚗 EV Owners | Find and navigate to nearest charging stations |
| ⛽ Petrol Station Owners | Evaluate conversion potential to EV charging |
┌─────────────────────────────────────────────────────────────────┐
│ 📍 Draw Any Area → 🔬 Analyze → 🎯 Get Optimal Locations │
└─────────────────────────────────────────────────────────────────┘
- Draw-to-Analyze: Draw custom polygons on any Kerala region for instant analysis
- 4-Layer Cost Analysis: Considers charging station proximity, population density, power grid connectivity, and EV adoption likelihood
- Heat Map Visualization: See favorability scores as GREEN (optimal) → YELLOW → RED
- Multi-Rank Results: Get top N alternative locations, not just "the best" one
- Real-time Navigation: Find and route to nearest EV or petrol station using GPS
| Metric | Current State | Ideal State |
|---|---|---|
| EV-to-Charger Ratio | 117:1 | 50:1 (IEA Standard) |
| Public Charging Stations | ~600 | 20,000+ (by 2030) |
| Registered EVs in Kerala | 70,000+ | 1,000,000 (2030 Target) |
| Rural Charger Coverage | 18% | 40%+ |
Key Challenges:
- ❌ Range Anxiety: #1 barrier to EV adoption
- ❌ Random Placement: Leads to underutilization in some areas, overcrowding in others
- ❌ Wasted Investment: ₹2,000+ Crore lost annually on suboptimal infrastructure
- ❌ Urban-Rural Gap: 2x disparity between urban and rural charging coverage
"Without data-driven planning, Kerala's 1 Million EVs by 2030 goal is at risk."
This tool transforms guesswork into science by:
- 📊 Analyzing 4 weighted factors for every grid cell in a region
- 🎯 Ranking locations from best to worst with transparent cost scores
- 🗺️ Visualizing 2,503 petrol stations as potential conversion sites
- 📈 Using 73 local body population zones for density calculations
- ⚡ Considering power substation proximity for grid connectivity
| Feature | Description |
|---|---|
| 🎨 Polygon Drawing | Draw any custom shape to analyze - no predefined boundaries |
| 📊 Multi-Layer Visualization | Toggle EV stations, petrol stations, density, substations, adoption rates |
| 🔥 Heat Map | See composite scores as color gradients (green = optimal) |
| 🏆 N-Rank Finder | Get top 1-10 locations ranked by cost score |
| 🔍 Region Browser | Filter and navigate between sub-locations in same rank |
| 📤 JSON Export | Export analysis results for GIS integration |
| Feature | Description |
|---|---|
| 📍 View All Stations | See 600+ EV charging stations and 2,503 petrol stations |
| 🧭 Find Nearest | Locate nearest EV or petrol station from your GPS location |
| 🗺️ Route Map | Embedded mini-map shows route with distance |
| 📱 Mobile Friendly | Responsive design works on any device |
View all 600+ EV charging stations (green markers) and 2,503 petrol stations (red markers) across Kerala at a glance. This visualization instantly reveals the infrastructure gap — where EVs can charge vs. where fossil fuel infrastructure dominates.
Draw any custom polygon on the map to trigger instant area analysis. The Grid View divides your selected region into analyzable cells while the Stats Panel displays key metrics — area size, charging stations count, EV vehicles estimate, EV penetration rate, and vehicle distribution.
Toggle to Heat Map View to see composite favorability scores as a color gradient:
- 🟢 Green (100%) — Highly favorable for new EV stations
- 🟡 Yellow (0%) — Neutral zones
- 🔴 Red (-100%) — Unfavorable (already saturated or low demand)
The algorithm considers 4 weighted factors: charging proximity, population density, substation distance, and EV adoption likelihood.
Click "Find Optimal" to compute the best locations for new charging stations. The system uses a divide-and-conquer algorithm to rank locations by cost score:
- Rank 1 (Green) — Best locations (lowest cost)
- Rank 2 (Blue) — Second-best alternatives
- Rank 3 (Purple) — Third-tier options
Each rank shows how many sub-locations share that score, enabling planners to choose from multiple equally-good sites.
For EV owners: Click "Navigate" to find the nearest charging station. The modal displays:
- Embedded mini-map with your route
- Station name and operator
- Distance to destination
- Option to "Find Another" if preferred
|
|
┌────────────────────┐ ┌────────────────────┐ ┌────────────────────┐
│ Next.js Frontend │────▶│ Next.js API Routes │────▶│ SQLite Database │
│ (Leaflet Maps) │ │ /api/stations │ │ (source.db) │
└────────────────────┘ └────────────────────┘ └────────────────────┘
│ │
│ ┌────────────────────┐ │
└────────────────▶│ Flask Backend │◀───────────────┘
│ :5000/api/... │
└────────────────────┘
| Requirement | Version | Check Command |
|---|---|---|
| Node.js | ≥ 18.0.0 | node --version |
| Python | ≥ 3.8 | python --version |
| npm/yarn/pnpm | Any | npm --version |
# 1. Clone the repository
git clone https://github.com/KenYeager/KERALA-MAP-ANALYSER.git
cd KERALA-MAP-ANALYSER
# 2. Install frontend dependencies
npm install
# 3. Install backend dependencies
cd backend
pip install flask flask-cors numpy scipy
cd ..npm run dev
# ✓ Ready at http://localhost:3000cd backend
python app.py
# ✓ Running on http://localhost:5000Open http://localhost:3000 in your browser. You should see:
- 🗺️ Map of Kerala centered at coordinates (10.8505, 76.2711)
- 🟢 Green markers showing EV charging stations
- 🔴 Red markers showing petrol stations
KERALA-MAP-ANALYSER/
├── 📂 backend/ # Flask Python backend
│ ├── app.py # Main Flask application (471 lines)
│ ├── evStationsLoader.py # EV station database queries
│ ├── petrolStationsLoader.py # Petrol station database queries
│ └── test_api.py # API testing utilities
│
├── 📂 cleaning/ # Data preparation & database
│ ├── source.db # SQLite database (≈7MB)
│ ├── ev-charging-station.csv # Raw EV station data (89K+ records)
│ ├── petrol.csv # Kerala petrol stations (2,503 records)
│ ├── kerala_local_body_indicators.csv # Population/density data
│ ├── import_to_sqlite.py # ETL script for importing data
│ ├── verify_db.py # Database verification script
│ └── clearner.py # Data cleaning utilities
│
├── 📂 public/ # Static assets
│
├── 📂 src/
│ ├── 📂 app/ # Next.js App Router
│ │ ├── layout.js # Root layout
│ │ ├── page.js # Homepage
│ │ ├── globals.css # Global styles + animations
│ │ └── 📂 api/ # API routes
│ │ ├── stations/route.js # GET /api/stations
│ │ ├── population_density/route.js
│ │ └── adoption_likelihood/route.js
│ │
│ ├── 📂 components/ # React components
│ │ ├── KeralMapAnalyzer.js # Main orchestrator (1,122 lines)
│ │ ├── Header.jsx # Top toolbar
│ │ ├── MapView.jsx # Leaflet map container
│ │ ├── StatsPanel.jsx # Analysis sidebar
│ │ ├── NavigationMenu.jsx # Find nearest station modal
│ │ ├── OptimalLocationModal.jsx # N-locations input dialog
│ │ ├── RegionSelector.jsx # Rank filter & navigation
│ │ └── 📂 stats/ # Stats panel cards
│ │ ├── AreaCard.jsx
│ │ ├── EVInfrastructureCard.jsx
│ │ ├── VehicleDistributionCard.jsx
│ │ ├── DemographicsCard.jsx
│ │ └── ...
│ │
│ └── 📂 utils/ # Utility modules
│ ├── mapUtils.js # Leaflet helpers, area calculations
│ ├── districtData.js # 14 Kerala districts metadata
│ ├── optimalLocationFinder.js # Visualization logic
│ ├── optimalLocationFinderAPI.js # Backend API client
│ ├── heatMapLayer.js # Heat map generation
│ ├── populationDensityLayer.js # Density overlay
│ ├── substationsLayer.js # Power substations
│ └── adoptionLikelihoodLayer.js # EV adoption scoring
│
├── 📄 TECHNICAL_DOCUMENTATION.md # Detailed algorithm documentation
├── 📄 package.json # Node.js dependencies
├── 📄 next.config.mjs # Next.js configuration
├── 📄 tailwind.config.js # Tailwind CSS configuration
└── 📄 README.md # This file
┌─────────────┐ ┌─────────────┐ ┌─────────────┐ ┌─────────────┐
│ 1. DRAW │────▶│ 2. ANALYZE │────▶│ 3. COMPUTE │────▶│ 4. DISPLAY │
│ Polygon │ │ Area │ │ Optimal │ │ Results │
└─────────────┘ └─────────────┘ └─────────────┘ └─────────────┘
│ │ │ │
▼ ▼ ▼ ▼
Click points Calculate: Flask backend Heat map +
on map to • Area (km²) processes grid Ranked regions
create polygon • Population with 4-factor with boundaries
• EV penetration cost algorithm
Each grid cell receives a cost score from -100 (excellent) to +100 (poor):
TOTAL_COST = (Proximity × 0.30) + (Density × 0.25) + (Substation × 0.25) + (Adoption × 0.20)
| Factor | Weight | Logic |
|---|---|---|
| Charging Proximity | 30% | PENALIZE cells near existing chargers (avoid clustering) |
| Population Density | 25% | FAVOR high-density areas (more users) |
| Substation Distance | 25% | FAVOR cells near power substations (cheaper grid connection) |
| Adoption Likelihood | 20% | FAVOR areas with high EV adoption propensity |
┌────────────────────────────────────────────────────────────┐
│ ALL GRID CELLS │
├────────────────┬───────────────────┬───────────────────────┤
│ 🟢 GREEN │ 🟡 YELLOW │ 🔴 RED │
│ cost ≤ -33 │ -33 < cost ≤ 33 │ cost > 33 │
│ (Favorable) │ (Neutral) │ (Unfavorable) │
├────────────────┴───────────────────┴───────────────────────┤
│ Process GREEN first → If N ranks found, STOP │
│ Only process YELLOW if more ranks needed │
│ Only process RED as last resort │
└────────────────────────────────────────────────────────────┘
The SQLite database (cleaning/source.db) contains 6 tables with Kerala infrastructure data:
-- ═══════════════════════════════════════════════════════════════════
-- TABLE 1: ev_stations (574 records)
-- Source: Open Charge Map API (filtered for Kerala)
-- Used by: Charging Proximity Cost Layer (30% weight)
-- ═══════════════════════════════════════════════════════════════════
ev_stations
├── id INTEGER PRIMARY KEY
├── latitude REAL -- GPS latitude
├── longitude REAL -- GPS longitude
├── status_code INTEGER -- Station operational status
├── access_code INTEGER -- Public/Private access
├── name TEXT -- Station name
├── operator TEXT -- Operating company
├── usage_type TEXT -- Type of usage
├── power_kw REAL -- Charging power in kW
└── connectors TEXT -- Connector types available
-- ═══════════════════════════════════════════════════════════════════
-- TABLE 2: petrol_stations (2,503 records)
-- Source: OpenStreetMap (Kerala bounding box)
-- Used by: Navigation feature (Find Nearest Station)
-- ═══════════════════════════════════════════════════════════════════
petrol_stations
├── id INTEGER PRIMARY KEY
├── latitude REAL -- GPS latitude
├── longitude REAL -- GPS longitude
├── name TEXT -- Station name
├── operator TEXT -- Operating company
├── brand TEXT -- Fuel brand (IOCL, BPCL, etc.)
├── city TEXT -- City/town location
├── phone TEXT -- Contact number
└── website TEXT -- Website URL
-- ═══════════════════════════════════════════════════════════════════
-- TABLE 3: population_density (73 records)
-- Source: Kerala Local Body Indicators (Census)
-- Used by: Population Density Cost Layer (25% weight)
-- ═══════════════════════════════════════════════════════════════════
population_density
├── latitude REAL -- Zone centroid latitude
├── longitude REAL -- Zone centroid longitude
├── population INTEGER -- Total population in zone
├── density_per_m2 REAL -- People per square meter
├── per_capita_income REAL -- Average income (₹)
└── area REAL -- Zone area in sq km
-- ═══════════════════════════════════════════════════════════════════
-- TABLE 4: adoption_likelihood (73 records)
-- Source: Derived from census + vehicle registration data
-- Used by: EV Adoption Likelihood Cost Layer (20% weight)
-- ═══════════════════════════════════════════════════════════════════
adoption_likelihood
├── latitude REAL -- Zone centroid latitude
├── longitude REAL -- Zone centroid longitude
├── population INTEGER -- Total population
├── ev_adoption_likelihood_score REAL -- 0-100 score (higher = more likely to adopt EV)
├── per_capita_income REAL -- Average income (₹)
└── area REAL -- Zone area in sq km
-- ═══════════════════════════════════════════════════════════════════
-- TABLE 5: SUBSTATIONS (116 records)
-- Source: Kerala State Electricity Board (KSEB)
-- Used by: Substation Proximity Cost Layer (25% weight)
-- ═══════════════════════════════════════════════════════════════════
SUBSTATIONS
├── Latitude REAL -- GPS latitude
├── Longitude REAL -- GPS longitude
└── Voltage_kV REAL -- Voltage capacity in kV
-- ═══════════════════════════════════════════════════════════════════
-- TABLE 6: EV_VEHICLES_PER_DISTRICT (67 records)
-- Source: Kerala Motor Vehicle Department
-- Used by: Stats Panel (Vehicle Distribution Card)
-- ═══════════════════════════════════════════════════════════════════
EV_VEHICLES_PER_DISTRICT
├── district TEXT -- District name
├── ev_count INTEGER -- Number of registered EVs
├── latitude REAL -- District centroid lat
└── longitude REAL -- District centroid lngThe algorithm evaluates each grid cell using 4 weighted cost factors. Lower total cost = more favorable for new EV charging station.
TOTAL_COST = (Proximity × 0.30) + (Density × 0.25) + (Substation × 0.25) + (Adoption × 0.20)
Purpose: PENALIZE cells near existing chargers to avoid clustering
| Parameter | Value | Description |
|---|---|---|
MAX_PENALTY_DISTANCE |
2.0 km | Beyond this, cells get negative cost (bonus) |
MAX_PENALTY_COST |
+100 | Cost at station location (worst) |
NEGATIVE_BONUS_CAP |
-50 | Maximum bonus for distant cells |
Algorithm:
if distance ≤ 2km:
penalty = (1 - (distance/2)²) × 100 // Quadratic decay
else:
bonus = min(50, (distance - 2) × 10) // Linear bonus, capped at -50
Database Fields Used: ev_stations.latitude, ev_stations.longitude
Purpose: FAVOR high-density areas (more potential EV users)
| Parameter | Value | Description |
|---|---|---|
WEIGHT |
-10,000 | Multiplier for density_per_m2 |
INFLUENCE_RADIUS |
Based on zone area | √(area/π) in km |
Algorithm:
For each density zone within influence radius:
contribution = density_per_m2 × WEIGHT × decay_factor
cell.cost += contribution
Database Fields Used: population_density.latitude, population_density.longitude, population_density.density_per_m2, population_density.area
Purpose: FAVOR cells near power substations (cheaper grid connection)
| Parameter | Value | Description |
|---|---|---|
MAX_BENEFIT_DISTANCE |
5.0 km | Influence radius |
MAX_BENEFIT_COST |
-50 | Cost reduction at substation (best) |
Algorithm:
if distance ≤ 5km:
benefit = (1 - distance/5) × (-50) × voltage_factor
cell.cost += benefit
Higher voltage substations provide stronger cost benefits.
Database Fields Used: SUBSTATIONS.Latitude, SUBSTATIONS.Longitude, SUBSTATIONS.Voltage_kV
Purpose: FAVOR areas with high EV adoption propensity
| Parameter | Value | Description |
|---|---|---|
INFLUENCE_RADIUS_KM |
3.0 km | Fixed influence radius |
MAX_COST_REDUCTION |
-20 | Maximum bonus for high adoption areas |
Algorithm:
For each adoption zone within 3km:
score = ev_adoption_likelihood_score (0-100 scale)
decay = 1 - (distance / 3000)²
cost_reduction = (score / 100) × (-20) × decay
cell.cost += cost_reduction
Database Fields Used: adoption_likelihood.latitude, adoption_likelihood.longitude, adoption_likelihood.ev_adoption_likelihood_score
| Cost Range | Color | Meaning |
|---|---|---|
| ≤ -33 | 🟢 GREEN | Highly favorable - optimal for new station |
| -33 to +33 | 🟡 YELLOW | Neutral - acceptable but not ideal |
| > +33 | 🔴 RED | Unfavorable - too close to existing infrastructure |
This project directly contributes to 4 United Nations Sustainable Development Goals:
SDG 7 Affordable & Clean Energy |
SDG 11 Sustainable Cities |
SDG 13 Climate Action |
SDG 17 Partnerships |
| SDG | Our Contribution | Measurable Impact |
|---|---|---|
| SDG 7 | Identify underserved areas for clean energy infrastructure | 40% coverage increase in rural areas |
| SDG 11 | Optimize station placement to reduce urban congestion | 30% reduction in average wait times |
| SDG 13 | Accelerate EV adoption by removing infrastructure barriers | Support Kerala's 1M EV target by 2030 |
| SDG 17 | Open-source platform for government + private sector collaboration | Multi-stakeholder data integration |
- ✅ First Indian state with dedicated EV policy (2019)
- ✅ Target: 1 million EVs by 2030
- ✅ This tool directly supports ANERT infrastructure planning
| Method | Endpoint | Description |
|---|---|---|
| GET | /api/stations?type=ev |
Fetch all EV charging stations |
| GET | /api/stations?type=petrol |
Fetch all petrol stations |
| POST | /api/population_density |
Get density data for bounds |
| POST | /api/adoption_likelihood |
Get adoption scores for bounds |
| Method | Endpoint | Description |
|---|---|---|
| GET | /api/health |
Backend health check |
| POST | /api/find-optimal-locations |
Compute optimal locations for polygon |
curl -X POST http://localhost:5000/api/find-optimal-locations \
-H "Content-Type: application/json" \
-d '{
"cells": [...],
"n": 5,
"minDistanceKm": 0.5
}'Response:
{
"success": true,
"executionTime": 0.245,
"cellsProcessed": 847,
"locations": [
{
"costRank": 1,
"cost": -55.00,
"subLocationCount": 3,
"subLocations": [...]
}
]
}Q: What data sources does this project use?
The project uses:
- EV Stations: OpenChargeMap API + OpenStreetMap
- Petrol Stations: OpenStreetMap Kerala extract (2,503 stations)
- Population Data: Kerala Local Body Indicators (73 zones)
- Adoption Likelihood: Computed from income + existing EV registrations
Q: Can I use this for areas outside Kerala?
Currently, the data is specific to Kerala. However, the codebase is modular—you can:
- Replace
source.dbwith your region's data - Update coordinate bounds in
districtData.js - Adjust cost calculation weights in
backend/app.py
Q: How accurate are the optimal location recommendations?
The algorithm considers 4 factors with configurable weights. Accuracy depends on:
- Data freshness (EV stations update regularly)
- Population density accuracy (2021 Census data)
- Substation data completeness
For production use, we recommend validating top recommendations with field surveys.
Q: Does this work offline?
Partially. The SQLite database works offline, but:
- Map tiles require internet (CartoDB Voyager tiles)
- Routing uses browser geolocation (requires network)
For fully offline use, consider caching map tiles with a tile server.
Q: How do I add more EV stations to the database?
cd cleaning
# Edit import_to_sqlite.py with your data source
python import_to_sqlite.py
python verify_db.py # Verify the importQ: What's the maximum polygon size I can analyze?
The system uses adaptive grid sizing:
- < 10 km²: 50m² cells (high precision)
- 10-50 km²: 100m² cells
- 50-100 km²: 200m² cells
-
100 km²: 500m² cells (warns user)
Very large polygons (> 500 km²) may take 10+ seconds to process.
Q: Can I export the analysis results?
Yes! Click the Export button in the header to download:
- Polygon coordinates
- All ranked locations with cost scores
- Cell-level data for GIS import (JSON format)
Contributions are welcome! Here's how you can help:
- Report Bugs: Open an issue describing the bug
- Request Features: Open an issue with the
enhancementlabel - Submit PRs: Fork, create a branch, make changes, submit PR
- Follow existing code style (ESLint + Prettier)
- Test changes locally before submitting
- Update documentation for new features
- Add comments for complex algorithms
This project is developed for the Asian Management Hackathon 2026.
- Kerala ANERT - EV policy framework and charging data
- OpenChargeMap - Global EV charging station database
- OpenStreetMap - Kerala petrol station data
- Leaflet.js - Open-source mapping library
- CartoDB - Map tile services (Voyager theme)