🛡️ CrimeIntel

AI-Driven Crime Analytics & Intelligence Platform

Transforming fragmented crime data into actionable intelligence for data-driven policing.

CrimeIntel unifies crime records, geospatial intelligence, analytical dashboards, hotspot detection, district intelligence, criminal-network analysis, and AI/ML-assisted insights in a secure decision-support platform for law-enforcement workflows.

Development / Demo Deployment:

Web Application: https://crime-intel-60079748823.development.catalystserverless.in/app/index.html
Backend API: https://crimeintel-backend-50044367664.development.catalystappsail.in

React + Vite · FastAPI · Supabase PostgreSQL · Supabase Auth + JWT · Multi-Tier Cache · Google Maps · Zoho Catalyst

⭐ Refined Prototype Phase — Upgrades & Enhancements

During the Refined Prototype Phase, CrimeIntel was enhanced beyond its initial functional prototype to improve deployment readiness, usability, accessibility, localization, privacy transparency, role-specific workflows, multi-tier caching performance, and operational validation.

🎨 UI/UX Refinement

Polished professional interface with consistent visual hierarchy tailored for police intelligence operations.
Smooth page transitions and loading states, including skeleton/loading placeholders for perceived performance.
Refined role-specific dashboards with distinct views for Field Officers, Intelligence Analysts, and Administrators.
Enhanced navigation bar, breadcrumbs, and standardized footer structure.
Back-to-top functionality and desktop-focused experience optimization.
Consistent Karnataka Police visual identity and styling system.
Comprehensive user settings/preferences experience with instant persistence.
Built-in Resources, Statutory Documents (PDF guides), and Support sections.

🗺️ Google Maps Upgrade

Replaced the initial Leaflet-based mapping layer with Google Maps (@vis.gl/react-google-maps).
Refined the geographic intelligence visualization experience to provide a familiar, high-performance, and integrated interface.
Integrated mapping seamlessly with the overall CrimeIntel design system.
Preserved all crime-intelligence and analytical capabilities (clusters, hotspots, density heatmaps, precinct boundaries).
Enhanced interactive marker selection and detailed case telemetry popups.

🌐 Language & Accessibility

Multilingual interface supporting English and Kannada (ಕನ್ನಡ).
Application UI localization for Kannada across navigation, dashboards, forms, telemetry labels, and alert messages.
Language preference is configurable through Settings and persists across sessions.
(Note: Localization primarily targets user-facing interface content; actual identifiers and backend data remain canonical.)

⚙️ User Preferences

Theme: Light / Dark mode toggle with instant theme application.
Formats: Configurable Date format and Time format preferences.
Landing: Default dashboard landing preference based on operational role.
Language: UI language selection (English / Kannada). These preferences persist in local storage to provide a tailored user experience across sessions.

🔐 Privacy, Consent & Security UX

Cookie Consent: Professional cookie consent banner for transparent data handling.
Privacy Policy & Terms of Service: Dedicated pages for transparency and governance trust.
Security Guidelines: Dedicated Security Guidelines documentation page.
Support: Integrated Contact Support experience with user-facing submission confirmation.

👥 Role-Based Experience Refinement

Tailored role-specific navigation and workflows for Field Officer, Intelligence Analyst, and Administrator.
Field Officer workflow focused on quick incident filtering, station beat cases, and local hotspot telemetry.
Intelligence Analyst workflow focused on cross-district comparisons, temporal trends, DBSCAN spatial clustering, CCRI risk scoring, and predictive forecasting.
Administrator workflow focused on security audit log queries, system status telemetry, and user management.

☁️ Zoho Catalyst Deployment

Zoho Catalyst AppSail: FastAPI/Python backend runs as an AppSail service with Python 3.10 runtime.
Catalyst Web Client Hosting: React 19 production build deployed to Catalyst Web Client Hosting.
Catalyst CLI: Integrated deployment workflows and automated packaging.

💾 Multi-Tier Caching & Backend Optimization

Multi-Tier Cache Architecture: Implemented a two-tier caching architecture combining a fast L1 in-memory LRU cache and an L2 Zoho Catalyst Cache segment.
L1 In-Memory LRU Cache: Thread-safe, bounded memory store (1,000 entries max) with per-item TTL expiration for sub-millisecond hot reads.
L2 Zoho Catalyst Cache: Shared BaaS cache segment for distributed persistence across AppSail worker instances.
Promotion & Fallback: Two-tier promotion on cache miss (L1 Miss $\rightarrow$ L2 Hit $\rightarrow$ Populate L1 $\rightarrow$ Return response). Cache failures fail safely directly to the repository layer without impacting API availability.
Single-Flight Coordination: Concurrency lock preventing cache stampedes during concurrent cache misses for identical keys.
Public Root Endpoint: Clean public GET / service status endpoint alongside health probes.

🚀 Scalability & Concurrency

Validated concurrency: The deployed AppSail backend handles concurrent simulated analytical workloads.
Verified operational resilience across core endpoints: Health, Readiness, Districts, District Intelligence, Analytics Summary, Hotspots, and Forecasting.

🧪 Testing & Validation

Automated backend test suite verifies 683 tests (683 passed, 92 deselected, 0 failed, 0 errors).
Comprehensive test coverage for JWT verification, server-side RBAC, multi-tier cache operations, rate limiting, audit logging, public root/health endpoints, and domain services.

📑 Table of Contents

Problem Statement
Solution Overview
Google Maps Integration
Key Features
System Architecture
Technology Stack
Data & Transparency
Security & Authentication
Project Structure
Getting Started
Configuration
API Overview
Testing & Reliability
Deployment
Production Extensions
Release Information
Team & Attribution
License

1. 🎯 Problem Statement

Law-enforcement agencies generate large volumes of operational information across FIRs, districts, police stations, arrests, chargesheets, victims, accused persons, and legal records. When these records remain fragmented across siloed files and reporting systems, extracting timely, actionable intelligence becomes difficult.

Challenge	Impact
Fragmented crime records	Information must be combined manually before meaningful strategic or field analysis can occur.
Limited analytical visibility	High-level trends, geographic hotspots, district variations, and repeat-offender links are difficult to identify quickly.
Reactive decision-making	Historical records exist, but converting them into proactive operational intelligence requires specialized processing.
Complex spatial and relational patterns	Geographic concentrations and multi-case criminal networks remain hidden within flat tabular records.

2. 💡 Solution Overview

CrimeIntel is an integrated AI-driven crime analytics and intelligence platform that converts structured police records into operational and strategic insights for data-driven policing.

Crime Dashboard · Geographic Crime Map · Hotspot Intelligence · District Intelligence · Trend Analytics · Network Analysis · AI/ML Insights · Secure Decision Support

CrimeIntel follows an API-first architecture. The modern React frontend communicates with secured FastAPI backend services. Supabase provides authentication and hosted PostgreSQL infrastructure, Zoho Catalyst provides cloud deployment (AppSail and Web Client Hosting), and Zoho Catalyst Cache provides distributed L2 caching.

User / Officer
      ↓
React + Vite Frontend (Catalyst Web Client)
      ↓  Bearer JWT (Supabase Auth)
FastAPI Backend (Zoho Catalyst AppSail)
      ↓
Multi-Tier Cache (L1 Memory → L2 Catalyst Cache)
      ↓ (Cache Miss)
Repository & ML Analytics Layer
      ↓
Supabase PostgreSQL Database

3. 🗺️ Google Maps Integration

During the Refined Prototype Phase, the initial mapping layer was upgraded to Google Maps (@vis.gl/react-google-maps) to provide a familiar, high-performance, and professional geographic visualization experience.

This integration aligns the spatial analysis workflow with modern law-enforcement UX standards while preserving all crime-intelligence and spatial telemetry features.

Prototype Evolution

Capability	Initial Prototype	Refined / Final Implementation
Mapping Technology	Leaflet	Google Maps (`@vis.gl/react-google-maps`)
Geographic Visualization	Basic	Enhanced Vector Maps + Satellite Layers
Crime Locations	Coordinate plotting	Interactive markers with rich metadata cards
Hotspot Detection	Bounded boxes	Dynamic DBSCAN clusters & intensity gradients
Density Heatmaps	Basic raster	Native high-density heatmap layers
Incident Clustering	Marker grouping	High-performance dynamic coordinate clustering
Role-Based Workflows	Unified map	Role-tailored Field Map & Intelligence Map

Geographic Capabilities

Crime Location Visualization: Precise coordinate plotting of incidents with crime category color coding.
Crime Hotspots: Visualizing geographic concentrations of recorded crime for patrol beat planning.
Heatmap Visualization: Rendering density intensity maps across urban and rural police precincts.
Cluster Visualization: Dynamically grouping dense incident records for clean map navigation.
Location-Based Telemetry: Interactive modal inspection with FIR summary, IPC sections, and station jurisdiction.

4. ✨ Key Features

📊 Interactive Executive & Role Dashboards

Consolidated crime KPIs, arrest rates, chargesheet distributions, temporal trends, district comparative metrics, and executive summaries tailored for Field Officers, Intelligence Analysts, and Station Commanders.

🗺️ Geospatial Crime Intelligence (Google Maps)

Interactive geographic visualization of incidents, precinct boundaries, dynamic DBSCAN clusters, density heatmaps, and spatial hotspot detection.

🔥 Crime Hotspot Intelligence

Identifies geographic concentrations of crime using spatial clustering algorithms to support targeted patrol deployment and resource allocation.

📈 Trend & Temporal Analytics

Explores crime patterns across time through multi-year timeline analysis, seasonal category trends, day/night temporal distributions, and cross-district comparative metrics.

🏙️ District & Station Intelligence Profiles

Comprehensive district-level profiles across all 31 Karnataka districts, including station rosters, crime category breakdowns, CCRI risk indicators, and recent FIR records.

🔗 Criminal Network Analysis

Builds deterministic relationship graphs from FIR-person relationships to reveal linked FIRs, person-case connections, co-accused associations, and repeat-offender clusters.

🧠 AI/ML Analytics & Forecasting

DBSCAN Spatial Hotspots: Pre-computed density-based spatial clustering of incident coordinates.
Composite Crime Risk Index (CCRI): Multi-factor risk scoring and ranking across police precincts.
Predictive Forecasting: Multi-day crime incident volume projections.

📤 Reporting, Export & Decision Support

Provides bounded operational CSV/PDF export, interactive visualizations, and structured analytical dossiers supporting evidence-based decision-making.

5. 🧩 System Architecture

                       +----------------------+
                       | Police / Analyst User|
                       +----------+-----------+
                                  |
                                HTTPS
                                  |
                                  v
+----------------------------------------------------------------+
|                     React + Vite Frontend                      |
| Dashboard | Field Map | Intelligence Map | District | Network  |
+---------------------------------+------------------------------+
                                  |
                              Bearer JWT
                                  |
                                  v
+----------------------------------------------------------------+
|                        FastAPI Backend                         |
|   Public Root (GET /)  |  Health Probes  |  Protected APIs     |
+---------------------------------+------------------------------+
                                  |
      +---------------------------+---------------------------+
      |                                                       |
      v                                                       v
+-----------------------------+             +-----------------------------+
|    L1 In-Memory LRU Cache   |             |    Authentication & RBAC    |
|   Thread-Safe · TTL Bounded |             |   JWT Verify · Claim Map    |
+--------------+--------------+             +-----------------------------+
               | (cache miss)
               v
+-----------------------------+
|   L2 Zoho Catalyst Cache    |
|   Segment Store · Fail-Safe |
+--------------+--------------+
               | (cache miss)
               v
+----------------------------------------------------------------+
|                   Repository & Service Layer                   |
|       PostgreSQL Repositories    |    ML & Analytics Engine    |
+---------------------------------+------------------------------+
                                  |
                                  v
+----------------------------------------------------------------+
|                       Supabase Platform                        |
|        Supabase Auth                 PostgreSQL Database       |
|        Session + JWT                 Production Persistence    |
+----------------------------------------------------------------+

      Cloud Deployment: Zoho Catalyst (AppSail + Web Client Hosting)

Multi-Tier Caching Architecture

L1 In-Memory LRU Cache: Fast in-process cache storing serializable response objects with per-item TTL expiration and capacity bounds (1,000 entries max).
L2 Zoho Catalyst Cache: Shared BaaS cache segment in Zoho Catalyst providing cross-worker cache persistence.
Promotion on Miss: When an L1 cache miss occurs, the system queries L2. On an L2 hit, the item is promoted to L1 for subsequent sub-millisecond retrieval.
Single-Flight Stampede Protection: Mutex coordination prevents redundant simultaneous backend queries for identical cache keys during heavy traffic.
Fail-Safe Operation: If L2 cache is unreachable or unconfigured, the application gracefully degrades to L1 and repository queries without throwing API errors.
Targeted Invalidation: Mutation operations trigger coordinated prefix and key invalidation across both L1 and L2 layers.

6. 🧰 Technology Stack

Layer	Technologies	Version / Details
Frontend	React 19, Vite 8, JavaScript/JSX	Modern component-driven UI
Styling & UI	Tailwind CSS, Framer Motion, Lucide React	Clean, professional dark/light police design
Mapping / GIS	Google Maps (`@vis.gl/react-google-maps`)	Coordinate plotting, heatmaps, clustering
Backend	Python 3.10, FastAPI, Uvicorn, Pydantic v2	High-performance asynchronous REST API
Database	PostgreSQL hosted on Supabase	Relational persistence, connection pooling
Authentication	Supabase Auth, JWT (PyJWT, Cryptography)	Cryptographic Bearer token verification
Authorization	Server-Side RBAC	Least-privilege role & permission mapping
Caching	Multi-Tier Cache (L1 LRU + L2 Zoho Catalyst)	In-memory TTL + Catalyst BaaS segment
Analytics & ML	Pandas, NumPy, Scikit-learn, XGBoost	DBSCAN clustering, CCRI scoring, forecasting
Network Intelligence	Graph analysis	FIR-person relationship link analysis
Testing	Pytest, TestClient, AnyIO	Automated testing suite (683 tests)
Cloud Hosting	Zoho Catalyst AppSail & Web Client Hosting	Managed serverless deployment

7. 📋 Data & Transparency

CrimeIntel adheres to strict data-integrity and privacy principles:

Approved Datasets as Source of Truth: Official crime records, FIR details, district boundaries, and station rosters serve as authoritative sources.
No Unattributed Synthetic Data: Synthetic or demo data is clearly separated and never presented as authoritative government records.
PII Exclusion in Analytical APIs: Person-level sensitive details (victim identities, full PII) are excluded from general analytical responses.
Architectural Separation: Persistence, business logic, caching, and presentation layers are decoupled via repository contracts.
Data Backend Support: Production persistence relies on PostgreSQL on Supabase, with an in-memory/CSV repository adapter available for offline unit testing.

8. 🔐 Security & Authentication

CrimeIntel implements a Zero-Trust, deny-by-default security architecture combining Supabase Auth for client sessions and FastAPI JWT validation for backend API enforcement.

User / Officer
      |
      v
React Frontend (Supabase Client)
      |
      | Sign in with departmental credentials
      v
Supabase Auth Service
      |
      | Returns Session + Access JWT
      v
Frontend API Client (`fetchAPI`)
      |
      | Authorization: Bearer <JWT>
      v
FastAPI Authentication Middleware & Dependencies (`get_current_identity`)
      |
      +---> Cryptographic JWT signature & expiry check (JWKS / Secret)
      +---> Server-side RBAC claim resolution (Admin / Analyst / Officer)
      +---> Route permission check (e.g. `dashboard.read`, `map.read`)
      |
      v
Protected CrimeIntel API Handlers

Security Controls

Public vs Protected Endpoints:
- Public Endpoints: Root service status (GET /) and health probes (/health, /health/live, /health/ready) are accessible without authentication.
- Protected Endpoints: All /api/v1/* routes require a valid Bearer JWT. Unauthenticated requests are rejected with HTTP 401 TOKEN_MISSING.
Cryptographic Verification: Validates token signatures against Supabase JWKS / JWT secrets, checking issuer, audience, and expiry claims.
Server-Side RBAC: Maps claims to validated roles (ADMIN, ANALYST, FIELD_OFFICER) with explicit granular permissions (dashboard.read, map.intelligence.read, audit.read, etc.).
Rate Limiting: In-process fixed-window rate limiter protecting cost-heavy endpoints (export, search, audit).
Security Headers: Enforces Cache-Control: no-store on authenticated API responses, X-Content-Type-Options: nosniff, and Referrer-Policy: strict-origin-when-cross-origin.
Security Audit Trail: Immutable append-only audit logging recording security events and denied access attempts.

9. 📂 Project Structure

Datathon/
├── backend/
│   ├── app/
│   │   ├── api/                 # FastAPI routes (dashboard, districts, maps, stations, analytics, auth)
│   │   ├── core/                # Auth dependencies, JWT verification, cache, audit, rate limit, config
│   │   ├── database/
│   │   │   ├── postgres/        # PostgreSQL implementation and connection pooling
│   │   │   ├── repositories/    # Repository contracts and CSV fallback adapters
│   │   │   └── ingest/          # Data ingestion utilities
│   │   ├── models/              # Pydantic and domain models
│   │   ├── schemas/             # Request/response schemas
│   │   ├── services/            # Business logic and intelligence services
│   │   └── main.py              # Application entry point, middleware, and route registration
│   ├── tests/                   # Pytest test suite (auth, rbac, cache, api, health, repositories)
│   ├── requirements.txt
│   ├── app-config.json          # Catalyst AppSail configuration
│   └── .env.example
├── frontend/
│   ├── src/
│   │   ├── components/          # Reusable UI elements, navigation, notifications, modals
│   │   ├── context/             # AuthContext (Supabase session), NotificationContext
│   │   ├── modules/             # Analytics, dashboard, district-intelligence, hotspot-detection, map, network
│   │   ├── services/            # Unified API client (`api.js`) and Supabase client (`supabase.js`)
│   │   ├── utils/               # PDF generation, formatters
│   │   └── App.jsx              # Main routing, role switching, and view controller
│   ├── public/                  # Public assets, icons, documentation resources
│   ├── package.json
│   ├── vite.config.js
│   └── .env.example
├── catalyst-web-client/         # Catalyst Web Client build output directory
├── docs/                        # Architecture, API specifications, and RBAC documentation
├── ml-engine/                   # ML models, clustering notebooks, and training pipelines
├── catalyst.json                # Catalyst project deployment configuration
└── README.md

10. 🚀 Getting Started

Prerequisites

Python 3.10+
Node.js 18+ and npm
Git
Access to Supabase project (URL and publishable/anon key)

1. Clone Repository

git clone https://github.com/Nirmal0804/Datathon.git
cd Datathon

2. Backend Setup

cd backend
python -m venv .venv

Windows PowerShell:

.\.venv\Scripts\Activate.ps1
pip install -r requirements.txt
Copy-Item .env.example .env
uvicorn app.main:app --reload --port 8000

Backend will be available at http://localhost:8000.

3. Frontend Setup

In a separate terminal:

cd frontend
npm install
Copy-Item .env.example .env
npm run dev

Frontend will be available at http://localhost:5173.

11. 🔧 Configuration

Backend Configuration (`backend/.env`)

APP_NAME=crime-analytics-backend
ENVIRONMENT=development
API_PREFIX=/api/v1
CORS_ORIGINS=["http://localhost:5173","http://localhost:3000"]

# Database Configuration
DATA_BACKEND=postgres
DATABASE_URL=postgresql://postgres:[PASSWORD]@db.[PROJECT-REF].supabase.co:5432/postgres

# Authentication Configuration
REQUIRE_AUTH=true
SUPABASE_PROJECT_REF=your-project-ref
SUPABASE_JWKS_URL=https://your-project-ref.supabase.co/auth/v1/.well-known/jwks.json
SUPABASE_JWT_ISSUER=https://your-project-ref.supabase.co/auth/v1
SUPABASE_JWT_AUDIENCE=authenticated

# L1 In-Memory Response Cache
CACHE_ENABLED=true
CACHE_TTL_SECONDS=600
CACHE_MAX_ENTRIES=1000

# L2 Zoho Catalyst Cache
CACHE_L2_ENABLED=true
CACHE_L2_SEGMENT_ID=your-catalyst-cache-segment-id
CACHE_L2_TTL_SECONDS=600

Frontend Configuration (`frontend/.env`)

VITE_API_BASE_URL=http://localhost:8000/api/v1
VITE_SUPABASE_URL=https://your-project-ref.supabase.co
VITE_SUPABASE_ANON_KEY=your-supabase-publishable-or-anon-key
VITE_GOOGLE_MAPS_API_KEY=your-google-maps-api-key

(Note: Never commit real passwords, secret keys, or service-role keys to source control.)

12. 📡 API Overview

Public Service & Health Probes

Method	Endpoint	Description	Auth Required
GET	`/`	Public root service status and message	No
GET	`/health`	Application status, database connection & cache statistics	No
GET	`/health/live`	Process liveness probe	No
GET	`/health/ready`	Readiness probe verifying PostgreSQL connection	No

Authentication

Method	Endpoint	Description	Auth Required
GET	`/api/v1/auth/me`	Returns server-verified identity and RBAC role	Yes (`Bearer`)

Executive Dashboard

Method	Endpoint	Description	Auth Required
GET	`/api/v1/dashboard/summary`	State-level KPIs, arrest rates, chargesheet distributions	Yes (`Bearer`)

Field Crime Map

Method	Endpoint	Description	Auth Required
GET	`/api/v1/map/field/cases`	Paginated FIR case records with geospatial filters	Yes (`Bearer`)
GET	`/api/v1/map/field/case/{fir_identifier}`	Detailed FIR record and case summary	Yes (`Bearer`)
GET	`/api/v1/map/field/filters`	Distinct filter options (districts, stages, crime types)	Yes (`Bearer`)
GET	`/api/v1/map/field/hotspots`	Beat-level spatial crime clusters	Yes (`Bearer`)

Intelligence Map & Analytics

Method	Endpoint	Description	Auth Required
GET	`/api/v1/map/intelligence/analytics`	High-level spatial intelligence metrics	Yes (`Bearer`)
GET	`/api/v1/map/intelligence/heatmap`	Spatial density coordinate points for heatmaps	Yes (`Bearer`)
GET	`/api/v1/map/intelligence/clusters`	Station-level incident clusters	Yes (`Bearer`)
GET	`/api/v1/map/intelligence/hotspots`	DBSCAN spatial hotspot centroids	Yes (`Bearer`)
GET	`/api/v1/map/intelligence/district-comparison`	Multi-district comparative crime metrics	Yes (`Bearer`)
GET	`/api/v1/map/intelligence/timeline`	Temporal crime incident trend data	Yes (`Bearer`)
GET	`/api/v1/map/intelligence/export`	Bounded operational CSV intelligence export	Yes (`Bearer`)

District & Station Intelligence

Method	Endpoint	Description	Auth Required
GET	`/api/v1/districts`	List all 31 districts with summary metrics	Yes (`Bearer`)
GET	`/api/v1/districts/{district_id}/intelligence`	Single-district comprehensive intelligence dossier	Yes (`Bearer`)
GET	`/api/v1/stations`	List police stations with optional district filtering	Yes (`Bearer`)
GET	`/api/v1/stations/{station_id}`	Detailed police precinct profile and metrics	Yes (`Bearer`)

Predictive & ML Analytics

Method	Endpoint	Description	Auth Required
GET	`/api/v1/analytics/summary`	Executive ML dashboard summary metrics	Yes (`Bearer`)
GET	`/api/v1/analytics/hotspots`	DBSCAN spatial cluster summaries	Yes (`Bearer`)
GET	`/api/v1/analytics/risk-scores`	Station-level CCRI risk ranks, scores, and tiers	Yes (`Bearer`)
GET	`/api/v1/analytics/forecast`	Daily crime incident volume forecast (1 to 30 days)	Yes (`Bearer`)

Criminal Network Analysis

Method	Endpoint	Description	Auth Required
GET	`/api/v1/network/graph`	Deterministic FIR-person relationship graph	Yes (`Bearer`)
GET	`/api/v1/network/entities/{type}/{id}`	Entity detail (FIR, person, station)	Yes (`Bearer`)
GET	`/api/v1/network/search`	Multi-entity cross-reference search	Yes (`Bearer`)

Security & Administration

Method	Endpoint	Description	Auth Required
GET	`/api/v1/admin/audit/events`	Query security audit event logs	Yes (`admin`)

13. ✅ Testing & Reliability

CrimeIntel is backed by a comprehensive automated test suite.

python -m pytest tests/ -k "not test_production_db_migration"

Verified Test Result:

===================== 683 passed, 92 deselected in 18.35s =====================

Reliability & Verification Coverage

Authentication & JWT Security: Valid token resolution, expired/malformed token rejection, JWKS validation, and algorithm-confusion protection.
Route Protection: Deny-by-default access verification across all protected routes.
Public Root & Probes: Verified GET /, /health, /health/live, and /health/ready responses.
Server-Side RBAC: Verified least-privilege role mappings and permission gates for Field Officers, Analysts, and Admins.
Multi-Tier Cache Operations: Unit and integration tests for L1 LRU memory store, L2 Catalyst Cache fallback, two-tier promotion, TTL eviction, and concurrency stampede coordination.
Rate Limiting: Verified fixed-window rate limiter on standard, search, export, and audit routes.
Domain Services: Tested dashboard metrics, district intelligence dossiers, station telemetry, DBSCAN hotspots, CCRI risk scoring, network graphs, and CSV data loader.

14. 📦 Deployment

CrimeIntel is deployed to Zoho Catalyst in a Development / Demo environment.

                     Zoho Catalyst Cloud
              +--------------------------------+
Browser ----> |  Catalyst Web Client Hosting   | (React 19 + Vite 8 App)
              +---------------+----------------+
                              |
                              v
              +--------------------------------+
              |     Catalyst AppSail (Python)  | (FastAPI Service)
              +---------------+----------------+
                              |
              +---------------+----------------+
              |                                |
              v                                v
     Zoho Catalyst Cache               Supabase Cloud
     (Distributed L2 Cache)      (Auth + PostgreSQL Database)

Deployed Endpoints (Development / Demo)

Frontend (Web Client): https://crime-intel-60079748823.development.catalystserverless.in/app/index.html
Backend (AppSail): https://crimeintel-backend-50044367664.development.catalystappsail.in

Deployment Commands

To deploy the AppSail backend:

catalyst deploy --only appsail:crimeintel-backend --ignore-scripts

To build and deploy the React frontend:

cd frontend
npm run build:catalyst
cd ..
catalyst deploy --only webclient

15. 🔮 Production Extensions

Implemented in Current Release

Role-Based Access Control (RBAC): Server-side role mapping (FIELD_OFFICER, ANALYST, ADMIN) with granular permissions.
Multi-Tier Response Caching: L1 in-process LRU cache combined with L2 Zoho Catalyst Cache segment.
Public Root Endpoint: Public GET / service response alongside health probes.
Google Maps GIS Integration: Advanced coordinate plotting, clustering, and heatmap layers.
AI/ML Predictive Analytics: DBSCAN spatial clustering, precinct CCRI risk scoring, and 30-day crime volume forecasting.
Security Audit Logging: Append-only security audit trail with query API.
Multi-Language Support: English and Kannada interface localization.
User Preference Persistence: Theme, date/time formatting, and default landing preferences.

Planned Future Extensions

Distributed Rate Limiting: Migrating in-process rate limiting to distributed store for multi-region active-active deployments.
Authoritative Boundary Ingestion: Integration with state GIS department boundary polygons.
Advanced Real-Time Telemetry: WebSocket integration for live officer location tracking and dispatch events.
Expanded Statutory PDF Reporting: Additional statutory compliance templates for automated court dossier generation.

16. 🏷️ Release Information

Current Version: v1.1.0
Release Status: Stable (Official Release)
Release Date: 6 September 2026
Release Channel: Final Datathon Release
Release Notes: RELEASE_NOTES.md
Documentation Asset: docs/CrimeIntel-Documentation.pdf
License: MIT License

17. 👥 Team & Attribution

Project: CrimeIntel — AI-Driven Crime Analytics & Intelligence Platform
Event: Karnataka Police Datathon 2026
Team Members: [To be completed]
Collaborator Names: [To be completed]
GitHub Usernames: [To be completed]
Acknowledgements: Karnataka State Police, Datathon Mentors & Organizers

18. 📄 License

This project is licensed under the MIT License — see the LICENSE file for details.

Note: The MIT License applies to the project's original source code and analytical models. Third-party packages, datasets, Google Maps services, Supabase infrastructure, and Zoho Catalyst cloud services remain subject to their respective terms and licenses.

🛡️ CrimeIntel

Crime Intelligence · Geospatial Analytics · Network Analysis · Decision Support

Transforming fragmented crime records into actionable intelligence.

Built for Datathon 2026

Name		Name	Last commit message	Last commit date
.agents/rules		.agents/rules
.github/workflows		.github/workflows
backend		backend
catalyst-web-client		catalyst-web-client
docs		docs
frontend		frontend
ml-engine		ml-engine
.gitignore		.gitignore
LICENSE		LICENSE
Procfile		Procfile
README.md		README.md
RELEASE_NOTES.md		RELEASE_NOTES.md
catalyst.json		catalyst.json

Latest commit

History

Folders and files

Repository files navigation

🛡️ CrimeIntel

AI-Driven Crime Analytics & Intelligence Platform

⭐ Refined Prototype Phase — Upgrades & Enhancements

🎨 UI/UX Refinement

🗺️ Google Maps Upgrade

🌐 Language & Accessibility

⚙️ User Preferences

🔐 Privacy, Consent & Security UX

👥 Role-Based Experience Refinement

☁️ Zoho Catalyst Deployment

💾 Multi-Tier Caching & Backend Optimization

🚀 Scalability & Concurrency

🧪 Testing & Validation

📑 Table of Contents

1. 🎯 Problem Statement

2. 💡 Solution Overview

3. 🗺️ Google Maps Integration

Prototype Evolution

Geographic Capabilities

4. ✨ Key Features

📊 Interactive Executive & Role Dashboards

🗺️ Geospatial Crime Intelligence (Google Maps)

🔥 Crime Hotspot Intelligence

📈 Trend & Temporal Analytics

🏙️ District & Station Intelligence Profiles

🔗 Criminal Network Analysis

🧠 AI/ML Analytics & Forecasting

📤 Reporting, Export & Decision Support

5. 🧩 System Architecture

Multi-Tier Caching Architecture

6. 🧰 Technology Stack

7. 📋 Data & Transparency

8. 🔐 Security & Authentication

Security Controls

9. 📂 Project Structure

10. 🚀 Getting Started

Prerequisites

1. Clone Repository

2. Backend Setup

3. Frontend Setup

11. 🔧 Configuration

Backend Configuration (backend/.env)

Frontend Configuration (frontend/.env)

12. 📡 API Overview

Public Service & Health Probes

Authentication

Executive Dashboard

Field Crime Map

Intelligence Map & Analytics

District & Station Intelligence

Predictive & ML Analytics

Criminal Network Analysis

Security & Administration

13. ✅ Testing & Reliability

Reliability & Verification Coverage

14. 📦 Deployment

Deployed Endpoints (Development / Demo)

Deployment Commands

15. 🔮 Production Extensions

Implemented in Current Release

Planned Future Extensions

16. 🏷️ Release Information

17. 👥 Team & Attribution

18. 📄 License

🛡️ CrimeIntel

Transforming fragmented crime records into actionable intelligence.

About

Topics

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages

Backend Configuration (`backend/.env`)

Frontend Configuration (`frontend/.env`)