A comparitive graph between actual load and predicted load
Flowchart
A comparision between Linear Regression and Moving Average , Linear Regression being more accurate.

Inspiration : With the increasing demand for electricity and frequent grid fluctuations, we wanted to explore how data and AI can help predict short-term energy usage. Our inspiration came from the need to build smarter, more reliable power systems that reduce wastage and outages.

What it does :

"PowerPredict" analyzes past energy consumption data and provides short-term forecasts (next few hours or day). These predictions can help utilities plan supply more effectively and ensure stable energy distribution.

How I built it :

Collected sample datasets of energy consumption in CSV format
- Preprocessed the data using "Python, Pandas, and NumPy"
- Applied "Scikit-learn regression models" to build forecasting models
- Visualized trends and predictions with "Matplotlib"
- Created a simple interactive web app using "Streamlit" to showcase results

Challenges I ran into :

Finding clean and reliable energy datasets
- Understanding time-series forecasting techniques as beginners
- Handling missing data and irregular time intervals
- Deploying the model into a simple app for demonstration

Accomplishments that I'm proud of :

Successfully building our first energy load forecasting model
- Creating clear visualizations that show actual vs predicted demand
- Learning to deploy a working prototype in a short time frame

What I learned :

Basics of time-series forecasting and machine learning
- Importance of data preprocessing and cleaning
- How to build and share interactive apps using Streamlit
- Teamwork and managing tasks under hackathon deadlines

What's next for PowerPredict :

Experimenting with advanced models like LSTM/GRU for more accurate forecasts
- Expanding to real-time data input and live predictions
- Adding renewable energy forecasting (like solar/wind)
- Deploying the app on cloud platforms for broader accessibility