⚡ WattWatch: Real-Time IoT Digital Twin & AI-Powered Smart Office Console

Winner-Grade Hackathon Entry // A production-ready, event-driven digital twin designed to monitor, analyze, and optimize office electrical consumption, preventing energy wastage through real-time web consoles and conversational AI Discord integrations.

---

🌟 Project Video & Live Demos

• Live Web Dashboard: https://wattwatch-f099.onrender.com
• Vercel Mirror: https://watt-watch-chi.vercel.app
• Discord Bot Username: Smart Circuit Bot (Active and listening 24/7)

---

📝 1. The Problem & Vision

In modern office spaces, electrical appliances (lights, fans, HVACs) are frequently left running overnight or in unoccupied rooms. This leads to:

1. Astronomical electricity bills that go unnoticed until the end of the month.
2. Severe environmental carbon footprints due to unoptimized energy draw.
3. Lack of immediate visibility of live consumption patterns for management.

WattWatch bridges this gap. By creating an Event-Driven Digital Twin of the office, any team member can see live device states, check current energy consumption (Watts/kWh), receive anomaly alerts, and query office telemetry directly from Discord using normal human language.

---

🗺️ 2. High-Level Event-Driven Architecture

WattWatch utilizes a single source of truth backed by a local SQLite database, sync'd instantly to clients via an HTTP Server-Sent Events (SSE) stream.

Data-Flow Walkthrough:

1. Simulation/Hardware Layer: The Python simulator (simulator.py) or physical ESP32 chips push telemetry and occupancy logs to Next.js API endpoints.
2. API & DB Unification: API endpoints write state updates (device toggle, occupancy change) directly into wattwatch.db (SQLite).
3. Real-Time Broadcast: The database layer triggers an SSE message (device_update, alert, usage_update, occupancy_update) over the persistent stream /api/stream.
4. Subscribers: The Next.js React Dashboard (Web UI) and the Python Discord Bot receive the update instantly, updating visual panels and sending proactive alerts without page refreshes.

---

🚀 3. Core Features

🖥️ A. Interactive Web Dashboard Console

• 2D Office Floorplan Blueprint: An interactive vector map showing three rooms (Drawing Room, Work Room 1, Work Room 2). Active lights emit a realistic soft glow filter, and active fans feature CSS keyframe spin animations.
• Live Device Control: Toggle all 15 devices (3 rooms, 2 fans and 3 lights per room) with a click. Writes write instantly to SQLite.
• Power Consumption Telemetry: Real-time total power draw (Watts), room-by-room breakdown, and daily accumulated energy consumption (kWh) calculator.
• Office Directory Panel: Renders contact cards of the registered employees pulling directly from SQLite.
• Active Alerts Log: Displays active warnings, allowing users to review and clear anomalies.

🤖 B. Conversational AI Discord Bot (discord_bot.py)

• Persistent Listener: Connects to the Next.js SSE stream and immediately posts rich warning embeds to a designated channel if devices are left ON in empty rooms.
• Gemini 2.5 Flash / Groq LLM Integration: Uses generative AI to translate raw telemetry values into friendly, professional conversational office reports.
• Command Set:
  • !status -> Returns conversational status of all rooms, active devices, and current occupancy counts.
  • !usage -> Summarizes live active load (W), daily energy used (kWh), and room draw.
  • !room <name> -> Returns live telemetry, occupant list, and active devices for a specific room (e.g. !room work1).
  • !occupants -> Lists registered office directory members.

---

🔌 4. Hardware ESP32 Circuit Specifications

For a physical deployment, we have designed a representative schematic detailing how each room is monitored.

• Actuation: The ESP32 controls power to the fans/lights via 5-channel Relay Modules (acting as isolator switches).
• Sensing: ACS712 Current Sensors are wired in series with each load to calculate RMS AC Current, translating it to real-time power consumption (Watts).

👉 Review the full hardware guide and circuit diagram: Hardware Circuit Schematic & Pin Mapping Specification.

---

👥 5. Mandatory Hackathon Dummy Dataset

The database contains seeded dummy profile directories exposed via the web directory panel and the Discord Bot (!occupants):

Name	Role	Email	Phone Number
Nafisa Rahman	Lead Developer	nafisa.rahman@yahoo.com	+8801812345678
Tanvir Hossain	System Admin	tanvir.hossain@yahoo.com	+8801912345678

---

⚙️ 6. Quick Setup & Run Instructions

Step 1: Clone and Install Dependencies

# Clone the repository
git clone https://github.com/Saobia3i/WattWatch.git
cd WattWatch

# Install Dashboard dependencies
npm install

# Install Python Bot & Simulator dependencies
pip install -r requirements.txt

Step 2: Configure Environment Variables (.env)

Create a .env file in the root folder (or copy .env.example):

# Next.js Server Configurations
NEXT_PUBLIC_API_URL=https://wattwatch-f099.onrender.com

# Discord Bot Configurations
DISCORD_BOT_TOKEN=your_discord_bot_token_here
DISCORD_ALERT_CHANNEL_ID=1430850944249233468

# Generative AI Key (Optional)
GEMINI_API_KEY=your_gemini_api_key_here

Step 3: Run the Web Dashboard

# Start Next.js development server
npm run dev

Step 4: Run the Python Simulator (Local Telemetry Feed)

python simulator.py

Step 5: Run the Discord Bot

python discord_bot.py

---

🔌 Hardware Simulation (Wokwi)

To simulate our office environment without physical hardware, we used Wokwi to build a virtual Arduino Mega setup. This simulates our 15 devices (fans and lights across 3 rooms), PIR motion sensors, and physical wall switches.

Wokwi Setup Instructions

1. Go to Wokwi Arduino Simulator and open a new Arduino Mega project.
2. Open the arduino/ folder in this repository.
3. Copy the contents of sketch.ino into the Wokwi code editor.
4. Copy the contents of diagram.json into the Wokwi diagram editor (this will automatically generate the wiring, sensors, and switches).
5. Click the green Play button to start the simulation.

How to Test and Verify Output

Once the simulation is running, you can manually interact with the virtual environment to test the logic:

• Triggering Sensors: Click on any of the PIR sensors and press "Simulate Motion" to mimic an employee walking into a room.
• Toggling Devices: Click the physical slide switches to manually turn individual fans and lights ON or OFF.
• Checking the Telemetry: Open the Serial Monitor at the bottom of the Wokwi screen.

You should immediately see live JSON strings printing every second, representing the current state of the office. It will look like this:

{"room":"drawing","occupants":1,"fan1":0,"fan2":1,"light1":1,"light2":0,"light3":0}

This JSON telemetry is what our Python/Node backend ingests to update the wattwatch.db database in real-time!

---

How the Hardware Logic Works

Our simulated Arduino Mega acts as the physical brain of the office, processing inputs and broadcasting state changes:

• Occupancy Tracking: We use PIR motion sensors mapped to entryways. The logic increments or decrements a peopleCount variable to track exactly how many employees are in a given room at any time.
• Manual Override Control: To prevent wasteful energy consumption (and to reflect real-world behavior), the devices do not turn on automatically when someone enters. Instead, the 15 devices (2 fans and 3 lights per room) strictly read the physical state of the manual slide switches.
• JSON Telemetry Generation: Every second, the Arduino's main loop reads the current switch states and sensor counts, packages them into a formatted JSON string, and pushes them out via the Serial Monitor. This acts as the raw data pipeline for our backend.

---

🤖 7. Discord Bot Command Cheat Sheet

Command	Action	Sample AI-Generated Output
!status	Checks status of all rooms	"Good afternoon team! Currently, the Work Room 1 has 1 light active (1 occupant), the Drawing Room has all devices off, and Work Room 2 is active with 1 fan running. The office is looking efficient!"
!usage	Returns current power draw	"System Check: We are pulling a total of 220 Watts right now. Today's total energy used is 4.872 kWh. Work Room 2 is drawing the most power (120W) due to the ceiling fan."
!room work1	Retrieves specific room stats	"Work Room 1: Currently occupied by 1 person. Fan 1 is OFF, Light 1 is ON. No energy wastage warnings."
!occupants	Shows employee contact list	"Here is the current registered office list: Nafisa Rahman (Lead Developer) & Tanvir Hossain (System Admin)."

---

🔒 8. Anomaly & Safety Rules

The system checks and issues alerts automatically for the following anomaly triggers:

1. After-Hours Activity: Any lights/fans left running outside 9 AM - 5 PM business hours.
2. Unoccupied Room Waste: Lights/fans running in a room where the sensor reads is_occupied = 0 for more than 15 minutes.
3. Continuous Run Warning: Any device running continuously for more than 2 hours without interruption.