SKIP TO CONTENT
POWERSENSE cover
FULL-STACK SOFTWARE DEVELOPMENTDemonstrated locally at Hackatron · No live demo

POWERSENSE

PowerSense — Smart Campus Energy Monitoring System

01

OVERVIEW

PowerSense is a full-stack intelligent campus energy-monitoring platform that monitors room-level electricity consumption, infers active electrical devices, detects abnormal energy patterns and provides actionable insights through an interactive dashboard.

02

PROBLEM STATEMENT

Institutional campuses waste significant energy: devices are left on after working hours, there is no room-level consumption visibility, and there is no automated anomaly detection, alerting or centralised energy analytics.

03

OBJECTIVES

  • ▸ Monitor room-level electricity consumption.
  • ▸ Infer active electrical devices.
  • ▸ Detect abnormal energy patterns.
  • ▸ Provide actionable insights through an interactive dashboard.
04

PROPOSED SOLUTION

A full-stack web application that reads live wattage per room through Socket.io, infers active devices with a subset-sum matching algorithm, flags anomalies with a machine-learning model, raises after-hours alerts, and calculates the energy, cost and CO₂ impact of the waste.

05

FEATURES

  • ▸ Real-time energy monitoring
  • ▸ Device load inference
  • ▸ Anomaly detection
  • ▸ After-hours energy monitoring
  • ▸ Incident alerts
  • ▸ Energy analytics
  • ▸ Financial impact estimation
  • ▸ CO₂ impact estimation
  • ▸ Campus infrastructure visualization
06

TECHNOLOGY STACK

ReactViteTailwind CSSNode.jsExpress.jsMongoDBSocket.ioPython (Flask)scikit-learnRechartsChart.jsFramer MotionElectron
07

SYSTEM ARCHITECTURE

Smart-meter readings reach an Express.js / Node.js REST API, are stored in MongoDB and pushed live to a React + Vite dashboard over Socket.io. A load-inference engine (subset-sum) and a Python Flask service (scikit-learn Isolation Forest) handle device matching and anomaly detection.

08

WORKFLOW

  • ▸ A smart meter sends real-time wattage data to the REST API.
  • ▸ The system matches the wattage against the device database using subset-sum.
  • ▸ The ML model analyses historical patterns to detect an anomaly.
  • ▸ The engine cross-checks room data and the configured working hours.
  • ▸ An alert is triggered and pushed to the admin dashboard via Socket.io.
  • ▸ Impact Forensics calculates the waste in kWh, ₹ and CO₂.
  • ▸ The administrator resolves the incident from the Anomalies page.
09

SCREENSHOTS

Live Intelligence dashboard — KPI cards and real-time load
Global Consumption Flow — live wattage across sectors
Infrastructure map — first floor, colour-coded room status
Infrastructure map — ground floor
Anomalies — incidents with matching device profiles
Impact Forensics — energy leak, financial deficit and CO₂ footprint
System configuration — institutions
System configuration — rooms
System configuration — devices and wattage profiles
10

CHALLENGES

  • ▸ Getting live data to show in real time. Solved by pushing meter readings to the dashboard instantly with Socket.io.
11

FUTURE IMPROVEMENTS

  • ▸ Connect real smart meters.
  • ▸ Send alerts by SMS.