System Requirements Document (SRD)

Project Name: retro-energy

1. Introduction

The retro-energy project aims to create a Python-based web application using Streamlit to forecast smart building energy consumption. This application will leverage a Random Forest Regressor model trained on synthetic data to provide real-time energy consumption predictions based on user-adjustable parameters. The project is designed to cater to professionals such as facility managers, researchers, and energy analysts, offering them an intuitive and interactive platform to explore energy trends and optimize building performance.

This document outlines the system requirements for retro-energy, ensuring clarity and alignment with the project's goals.

2. System Overview

The retro-energy system will consist of the following core components:

1. Machine Learning Model: A Random Forest Regressor trained on a synthetic dataset with features such as outdoor temperature, occupancy, hour of the day, and humidity.
2. User Interface: A Streamlit-based web application with a clean, modern "Smart Building" theme.
3. Dynamic Interactivity: Real-time updates to predictions and visualizations as users adjust input sliders.
4. Visualization Features:
   • A dynamic Plotly line chart displaying predicted energy consumption trends.
   • Individual graphs for each regression module, showing the relationship between input variables and energy consumption.
5. User Experience: A seamless, responsive interface designed for professionals in the energy and building management sectors.

3. Functional Requirements

• As a User, I should be able to adjust input sliders for outdoor temperature, occupancy, hour of the day, and humidity in real-time.
• As a User, I should see the predicted energy consumption (in kW) prominently displayed on the screen.
• As a User, I should see a dynamic Plotly line chart that updates as I adjust the sliders, showing the predicted energy consumption trend.
• As a User, I should see individual graphs for each regression module (outdoor temperature, occupancy, hour of the day, and humidity) to understand their impact on energy consumption.
• As a User, I should be able to compare the current prediction with the "average" building performance.

4. User Personas

1. Facility Manager

   • Role: Oversees building operations and energy efficiency.
   • Goals: Optimize energy usage, reduce costs, and ensure sustainability.
   • Needs: Real-time insights, intuitive controls, and actionable data.

2. Researcher

   • Role: Studies energy consumption patterns and building performance.
   • Goals: Analyze trends, validate hypotheses, and publish findings.
   • Needs: Detailed visualizations, transparency in model behavior, and exportable data.

3. Energy Analyst

   • Role: Evaluates energy consumption for multiple buildings or systems.
   • Goals: Identify inefficiencies, forecast demand, and recommend improvements.
   • Needs: Accurate predictions, comparative analytics, and feature importance insights.

5. Visuals Colors and Theme

Color Palette for retro-energy:

• Background: #F4F9F9 (Soft Mint Green)
• Surface: #E8F1F2 (Light Aqua)
• Text: #2C3E50 (Deep Navy Blue)
• Accent: #16A085 (Emerald Green)
• Muted Tones: #BDC3C7 (Soft Gray)

This palette reflects a clean, modern, and professional aesthetic, aligning with the "Smart Building" theme.

6. Signature Design Concept

Interactive Energy Dashboard with Modular Insights

The homepage of retro-energy will feature a dynamic, modular dashboard that feels alive and engaging.

1. Hero Section:

   • A sleek, animated energy waveform graphic flows across the top of the page, representing real-time energy dynamics.
   • The title "retro-energy" appears with a subtle glow effect, reinforcing the futuristic theme.

2. Interactive Modules:

   • The dashboard is divided into four collapsible panels, each representing a regression module (outdoor temperature, occupancy, hour of the day, and humidity).
   • Each panel includes a Plotly graph that updates dynamically as users adjust the corresponding slider.

3. Micro-Interactions:

   • Sliders animate with a smooth "snap" effect when adjusted.
   • Hovering over graphs reveals tooltips with detailed data points.

4. Real-Time Feedback:

   • The predicted energy consumption is displayed prominently in the center of the dashboard, with a glowing ring animation that pulses gently to indicate real-time updates.

This design ensures that users are not only informed but also engaged, making energy forecasting an intuitive and visually appealing experience.

7. Non-Functional Requirements

• Performance: The system should provide real-time predictions with minimal latency (<1 second).
• Scalability: The application should support concurrent users without performance degradation.
• Usability: The interface should be intuitive and accessible, requiring no prior technical expertise.
• Reliability: The system should handle edge cases (e.g., extreme input values) gracefully.
• Security: User interactions and data should be securely processed and stored.

8. Tech Stack

• Frontend:

  • Streamlit for web application development.

• Backend:

  • Python for model integration and API logic.

• Database:

  • MySQL for storing user interaction data and average building performance metrics.

• AI Models:

  • Random Forest Regressor for energy consumption prediction.

• Visualization Tools:

  • Plotly for dynamic graph generation.

• Orchestration:

  • Docker for containerization.

9. Assumptions and Constraints

• Assumptions:

  • Users will have access to a modern web browser to interact with the application.
  • The synthetic dataset will accurately represent real-world energy consumption patterns.

• Constraints:

  • The application must run on a single server instance initially.
  • The system will not include advanced user authentication in the first release.

10. Glossary

• Regression Module: A component of the model that predicts energy consumption based on a specific input variable.
• Synthetic Dataset: A dataset generated artificially to simulate real-world data.
• Plotly: A graphing library used for creating interactive visualizations.
• Random Forest Regressor: A machine learning algorithm used for regression tasks.
• Streamlit: A Python library for building interactive web applications.