Back to Srishti's profile
13views

EstateIQ

EstateIQ uses a hybrid Ubuntu + AWS architecture, where Ubuntu handles local data processing and monitoring, while AWS provides cloud storage, AI/ML analytics,

Srishti Kumari
  • Problem, approach and result explained
  • Expert reviewed· 6 reviews
EstateIQ

—

92

—

96

—

93

Overview

The problem

EstateIQ addresses the challenges faced by colleges, hospitals, government institutions, and large facilities in monitoring and managing energy and water consumption. Traditional systems rely on manual monitoring and historical data, making it difficult to detect anomalies, identify inefficiencies, and predict future requirements. This leads to resource wastage, increased operational costs, and delayed decision-making. EstateIQ solves this through an AI-powered IoT platform that enables real-time monitoring, anomaly detection, consumption forecasting, root-cause analysis, and intelligent recommendations for smarter and more sustainable facility management.

The approach

EstateIQ is built as an AI-powered facility intelligence platform that integrates IoT-based monitoring, cloud infrastructure, data analytics, and machine learning to help institutions make data-driven operational decisions.

The hardware layer uses ESP32-based IoT nodes connected to sensors for electricity consumption, water flow, temperature, humidity, and air quality monitoring. Sensor readings are collected at configurable intervals and transmitted through Wi-Fi using MQTT/HTTP communication.

The backend architecture uses an Ubuntu-based processing environment and AWS services for scalable data handling. AWS IoT Core manages device communication, AWS Lambda handles event-driven processing, and Amazon S3 provides cost-efficient storage for historical readings. MongoDB stores structured operational data for dashboard queries and analysis.

The application dashboard is developed using the MERN stack, providing real-time monitoring, historical trends, anomaly alerts, and facility-level performance insights. Machine learning models using Scikit-learn and forecasting techniques help identify unusual consumption patterns, predict short-term demand, and generate actionable recommendations.

One of the most challenging parts was designing reliable monitoring for multiple facilities while handling inconsistent sensor readings, communication failures, and abnormal values. We addressed this through configurable sampling intervals, sensor health checks, anomaly detection, and alert escalation mechanisms.

We deliberately prioritized electricity, water, air quality, and overall facility performance in the initial implementation. Advanced integrations and large-scale deployment across multiple institutions were kept outside the initial prototype scope to maintain a practical and achievable solution.

The result

EstateIQ demonstrates how raw facility data can be transformed into meaningful operational intelligence through a unified monitoring and decision-support dashboard.

The prototype brings together multiple monitoring parameters, allowing administrators to observe electricity consumption, water usage, environmental conditions, and facility performance from a centralized interface.

The platform demonstrates the following capabilities:

  • Centralized monitoring of multiple buildings and connected IoT parameters.
  • Live sensor-data collection and visualization through the dashboard.
  • Detection of abnormal readings and unusual consumption patterns.
  • Sensor health monitoring to identify communication failures and potentially frozen readings.
  • Short-term electricity demand forecasting.
  • AI-assisted insights explaining possible causes of abnormal consumption.
  • What-If simulation to estimate potential energy savings from operational changes.
  • Actionable recommendations to support better resource utilization.
  • Historical data analysis for identifying consumption trends and performance variations.

The system uses sensor readings and controlled test datasets to demonstrate its monitoring, analytics, and prediction workflows. Results are evaluated by comparing detected anomalies against expected operating patterns, comparing predicted consumption with actual or reference readings, and checking whether alerts are generated under defined conditions.

The prototype establishes the complete concept of moving from sensor-level data collection to centralized monitoring, intelligent analysis, and actionable recommendations.

Our key outcome is demonstrating that EstateIQ is not simply a monitoring dashboard; it is a decision-intelligence system that helps administrators understand problems, anticipate operational requirements, and take informed action.

Reflection

During the development of EstateIQ, we learned that building a reliable facility intelligence platform involves much more than connecting sensors and displaying readings. Data accuracy, communication reliability, scalability, and meaningful interpretation are equally important.

One of the biggest challenges was ensuring that sensor readings accurately represented real-world conditions. Low-cost sensors can produce inconsistent readings, and network interruptions can affect data transmission. We recognized the importance of sensor calibration, data validation, device health monitoring, and reliable communication protocols.

If we were to rebuild the project, we would prioritize hardware calibration and long-duration testing from the beginning. We would also introduce more advanced device management, automated recovery mechanisms, and stronger validation of machine learning predictions.

Some features still require further development and real-world validation, including large-scale deployment, long-term forecasting accuracy, automated recommendation execution, and integration with industrial-grade sensors and existing building management systems.

We also learned that AI recommendations must be explainable and practical. Administrators should understand not only what the system recommends but also why the recommendation was generated and what impact it could have.

Our biggest learning was that sustainability is not achieved by collecting more data; it is achieved by converting operational data into measurable decisions and improvements.

EstateIQ is designed with this principle at its core, and our next step is to move from prototype validation towards a reliable, scalable, real-world facility intelligence solution.

Submitted to Sustainable Facility & Estate Intelligence

Your idea in one line

We have created EstateIQ which is a dashboard for facility teams or management in corporate ,institution ,hospital or any PSU campuses .It focuses on energy conservation and on achieving sustainability. It analyses , processes , recommends better sustainable ways for energy wastages .It puts electricity and water data in one place, ranks which problem to fix first and why, and shows what a fix would save before anyone makes it.

What will be working by the finale?

By the finale, we aim to deliver a functional EstateIQ decision-intelligence platform with an integrated facility dashboard, AI-driven anomaly detection, energy-consumption forecasting, explainable alerts, and our proposed Elite Algorithm for prioritizing operational issues. We will also demonstrate a What-If Simulator to evaluate potential corrective actions, along with actionable recommendations and decision tracking. Our objective is to demonstrate an end-to-end workflow that transforms facility data into actionable insights for resource optimization and sustainable operations.

Tech stack and hardware you plan to use

Backend: Python and FastAPI. Dashboard: Streamlit, with Plotly for charts. Forecasting: XGBoost, LightGBM, CatBoost and Prophet, and SHAP to explain each alert. Hardware: a Raspberry Pi with a three-phase energy meter and current transformer for electricity, and a water flow meter and tank level sensor for water. A local Ubuntu server runs everything offline, and AWS gives remote access. Login uses JWT with role-based access, and data is encrypted.

A diagram, sketch or mock-up of your idea
A diagram, sketch or mock-up of your idea: EstateIQ
What have you built or tested so far?

So far, we have developed the EstateIQ dashboard and core platform structure, worked on the AI and analytics modules, and designed our proposed Elite Algorithm to integrate predictive signals, anomaly detection, explainability, and decision prioritization. We have also developed the product demonstration workflow, including facility monitoring, alerts, forecasting, and What-If analysis. Our current focus is validating the end-to-end integration, evaluating model performance, and ensuring the recommendations are reliable and actionable

Pitch deck (PDF)
Pitch deck (PDF)PDF
What's working now?

We have developed our Elite algorithm , which is a collection of multiple open source algorithms . Instead of calculating each module separately by using the Elite algorithm , we can calculate them at once which makes everything faster, we have implemented AI in our project which help us to rectify anomalies gives prediction , forecasts on previous data and what if simulators which helps us to make better decisions , and cost effective optimizations in different scenarios. We have created a dashboard for electricity , water , air ,waste management which monitors the real time data and if any anomaly gets detected it sends the alert in telegram chatbot . We lag real time data due to not being able to use industry level sensors because of affordability issues. We tried to work on traffic but didn't get any such solution which would have been a better solution than the existing solutions and also would be helpful in day to day life.

Deck or report (PDF)
Deck or report (PDF)PDF
A photo or screenshot
A photo or screenshot: EstateIQ

AI tools used

ChatGPTClaude

Chatgpt and Claude for scripting

Gallery

5