Welcome to our AI-powered solution developed for Smart India Hackathon 2024, aimed at solving one of the real-world problem statements using predictive intelligence.
This project leverages historical data and advanced machine learning techniques to predict optimal time slots for processes, appointments, or sessions – ensuring efficiency, reduced wait times, and smart decision-making.
- Manual scheduling is inefficient.
- Peak hours create congestion.
- Lack of data-driven decisions leads to poor resource use.
Our model fixes this by predicting time slots using intelligent data analysis and forecasting techniques, which can easily be integrated into real-world appointment systems.
- 📊 Predicts time slots based on historical usage.
- 🤖 AI model trained on real data.
- 💡 Fast API for predictions.
- 🧠 Supports scalability for large datasets.
- ⚙️ Easily customizable and deployable.
- Python for ML model development
- Pandas, NumPy, Scikit-learn for data processing and ML
- FastAPI for API endpoints
- Git + GitHub for version control
- Jupyter Notebook for model experimentation
SYSTEM ARCHITECTURE & DATA FLOW
[START]
|
V
+-------------------------+
| User-Facing Web Portal |
+-------------------------+
|
+-------------------+-------------------+
| |
// --- SENDER's JOURNEY --- // --- RECIPIENT's JOURNEY ---
| |
V V
+----------------------------+ +-----------------------------+
| 1. Sender Enters Ref. No. | | 1a. Recipient Enters Ref. No|
+----------------------------+ +-----------------------------+
| |
V V
+----------------------------+ +-----------------------------+
| 2. AI Model Predicts Time | | 2a. Track & View Time |
+----------------------------+ +-----------------------------+
| |
V V
< 3. Sender Satisfied? > < 3a. Recipient Satisfied? >
| | | |
YES | | NO YES | | NO
| | | |
| +---------------------+ | +----------------------+
| | Proposes New Time | | | Proposes New Time |
| +---------------------+ | +----------------------+
| | | |
| V | V
| < AI Re-evaluates > | < AI Re-evaluates >
| | | |
+--------------+----------------------)---------------+--------------+
|
V
+-----------------------------------------+
| 4. SCHEDULING & BACKEND CONFIRMATION |
+-----------------------------------------+
|
<<============== Writes & Reads =============>>
+-----------------------------------------------------+
| D A T A B A S E |
| (User Info, Historical Success Rates, Preferences) |
+-----------------------------------------------------+
<<===============================================>>
|
V
// --- PORTAL OFFICE's WORKFLOW ---
+-----------------------------------------+
| 5. View Confirmed Schedule on Dashboard|
+-----------------------------------------+
|
V
6. OPTIMIZE ROUTE (optimize_route.py using Google OR-Tools)
|
V
+-------------------------+
| 7. Assign Route to Staff |
+-------------------------+
|
V
< Can Delivery Be Made? >
| |
YES | | NO
| |
V +-----> (Reschedule Within Range)
+---------------------+
| 8. DELIVERY SUCCESSFUL|
+---------------------+
|
V
+---------------------+
| 9. Collect User Rating|-----> (Update Database w/ New Data)
+---------------------+
|
V
[STOP]
git clone https://github.com/ashmitasenroy/SIH---AI-based-time-slot-prediction-model.git
cd SIH---AI-based-time-slot-prediction-modelpip install -r requirements.txt- To test the prediction logic:
python model.py- To run the FastAPI app:
uvicorn app:app --reloadVisit: http://127.0.0.1:8000/docs for Swagger UI.
Ritusree Das Mohak Das Rudranil Choudhary Surya Pratap Verma Anisha Singh (@anisha-singh-2004) Ashmita Sen Roy (@ashmitasenroy)
Thanks to SIH mentors and coordinators for their support!
This project is under the MIT License. See the LICENSE file for more info. © 2025 Ashmita Sen Roy
Feel free to raise issues or contribute via pull requests.