Inspiration
We wanted to understand why electricity bills vary so much by city, building type, and season. With rising energy costs, we were inspired to build a tool that makes power usage and monthly costs easier to predict and understand.
What it does
Energy Calculator predicts monthly electricity usage, cost per kWh, and estimated cost per household for a given city, year, and building type. It compares the same months across different years to capture seasonal trends and outputs results in a format that can be used directly by a Java app.
How we built it
We cleaned and processed a real-world energy dataset using Python and pandas, then trained linear regression models to predict kWh usage, cost per kWh, and temperature. We calculated household costs using population-based estimates and building-type adjustments, and connected everything to a Java frontend using JSON output.
Challenges we ran into
Handling messy data was difficult, especially inconsistent month formats and missing values. We also had to fix logical errors in the household cost formula and resolve pandas warnings caused by dataframe slicing. Making sure predictions compared the same months across years was another key challenge.
Accomplishments that we're proud of
We built a full end-to-end system that cleans data, trains models, makes predictions, and integrates with another programming language. We also corrected flawed cost calculations to produce realistic household-level estimates.
What we learned
We learned how linear regression works in practice, how important data cleaning is, and how small mistakes in formulas can cause big errors. We also gained experience debugging, working with pandas, and integrating Python with Java.
What's next for Energy Calculator
Next, we want to add more advanced models, support more cities and datasets, and improve accuracy by including factors like renewable energy usage and weather patterns.
Log in or sign up for Devpost to join the conversation.