FIRE is an intelligent AI-driven routing and analytics system designed to process, analyze, and geographically distribute customer support tickets and client requests. It features a high-density, professional "Bloomberg Terminal V1" aesthetic.
Jupyter Notebook + Kaggle Notebook
- Tickets: 186
- Managers: 51
- Business Units: 15
Охват клиентов менеджерами: 100.00% (sanity-check, что все клиенты были распределены)
| assigned_office | min | max | mean | num_managers |
|---|---|---|---|---|
| астана | 1 | 2 | 1.42857 | 7 |
| алматы | 3 | 16 | 10.1667 | 6 |
| шымкент | 3 | 15 | 6.6 | 5 |
| актобе | 1 | 3 | 1.66667 | 3 |
| усть-каменогорск | 2 | 20 | 8 | 3 |
| костанай | 2 | 9 | 6 | 3 |
| актау | 2 | 3 | 2.66667 | 3 |
| павлодар | 7 | 9 | 8 | 3 |
| кокшетау | 9 | 20 | 14.5 | 2 |
| атырау | 3 | 6 | 4.5 | 2 |
| караганда | 8 | 11 | 9.5 | 2 |
| петропавловск | 1 | 2 | 1.5 | 2 |
| кызылорда | 1 | 1 | 1 | 1 |
| уральск | 3 | 3 | 3 | 1 |
| тараз | 1 | 1 | 1 | 1 |
| Офис | Среднее расстояние, км | |
|---|---|---|
| 0 | актобе | 722.997 |
| 1 | кызылорда | 369.211 |
| 2 | актау | 300.284 |
| 3 | павлодар | 197.182 |
| 4 | алматы | 183.992 |
| 5 | кокшетау | 160.682 |
| 6 | уральск | 155.913 |
| 7 | петропавловск | 149.825 |
| 8 | костанай | 124.41 |
| 9 | караганда | 105.426 |
| 10 | атырау | 85.3087 |
| 11 | шымкент | 75.7876 |
| 12 | усть-каменогорск | 66.2079 |
| 13 | астана | 3.21285 |
| 14 | тараз | 1.55725 |
The frontend is a data-dense, real-time analytics dashboard built for professional asset managers.
- Framework: React 18 + Vite
- Styling: Vanilla CSS + Tailwind CSS (configured for a dark, high-contrast terminal aesthetic)
- Data Visualization: Recharts (for analytics and KPI metrics)
- Interactive Maps: Leaflet & React-Leaflet (for geographical ticket distribution and manager load balancing)
- Icons: Lucide React
The backend is a high-performance Python server that handles AI enrichment and database ORM.
- Framework: FastAPI (with Uvicorn ASGI server)
- Database ORM: SQLAlchemy
- Data Validation: Pydantic
- AI/LLM Engine: Google Gemini (Gemini 1.5 Pro) for natural language understanding, tone analysis, and priority scoring.
- Provider: Serverless PostgreSQL on Neon Database
- Schema: Core business data (read-only Cyrillic tables) joined with an AI enrichment layer (
ticket_enrichmentandoffice_enrichment) for non-destructive metadata augmentation.
To run the FIRE application locally, you need to run both the frontend and backend servers simultaneously.
Ensure you have a Neon Database connection string. Create a .env file in the backend/ directory and add your database URL, 2GIS API key, and Groq API key:
NEON_DATABASE_URL=postgresql://user:[email protected]/dbname?sslmode=require
API_KEY_2GIS = get from: https://dev.2gis.ru/en/api
GROQ_API_KEY = get from: https://console.groq.com/keys
GEMINI_API_KEY = get from: https://aistudio.google.com Navigate to the backend directory, install the dependencies, and start the FastAPI server:
cd backend
# Create and activate a virtual environment (recommended)
python3 -m venv venv
source venv/bin/activate
# Install dependencies
pip install -r requirements.txt
# Start the Uvicorn server
uvicorn main:app --reloadThe backend API will be running at http://127.0.0.1:8000.
Open a new terminal window, navigate to the frontend directory, install the packages, and start the Vite development server:
cd frontend
# Install dependencies
npm install
# Start the development server
npm run devThe frontend UI will be running at http://localhost:5173/.
/backend/main.py: The entry point for the FastAPI application./backend/models.py: SQLAlchemy ORM models matching the Neon database schema./backend/routers/: API endpoints for analytics, tickets, managers, and the AI assistant./backend/services/ai_service.py: Integration with Google Gemini for text analysis./frontend/src/index.css: Core terminal theme CSS variables./frontend/tailwind.config.js: Custom color palette definitions (e.g.,bg-primary,accent-gold)./frontend/src/pages/: React components for Dashboard, Analytics, Tickets, Managers, User Synthetic Data generation, and Assistant UI.