solar-calculator-models

KO | EN

gitlite — search

#javascript #python #react #ai #llm #android #rust #java #windows #cli #vue #pytorch #mcp #open-source

solar-calculator-models

(★ 13)

technical report and modeling code used for solarcalculator.xyz

solar-calculator-models Latest Version Download

Download Latest Version (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Solar Output and Cost Prediction
This repo contains the code used to build machine learning models for solarcalculator.xyz.  For a non-technical report on the work contained in this repo, head over to my article on Medium: https://medium.com/@evanbaker/.  Code for the Django app is contained in a seperate repo.

For a non-technical writeup on this project, head over to my <a href=https://medium.com/@evanbaker/predicting-solar-energy-production-with-machine-learning-19fcab295e58>article on Medium</a>

### Goal
The aim of this project is to use machine learning to build two predictive models, one that predicts annual energy generation, and another that predicts installation cost, in order to ultimately create a user-friendly web service that allows a user to understand what their cost would be of installing solar, how long it would take to pay off, and how much they should expect to subsequently save on their energy bill every year.  Using a random forest model, it is possible to predict expected annual return on energy generated from solar to a median accuracy of +/- $15.00, a 75% increase in accuracy over predictions generated by the National Renewable Energy Laboratory (NREL).

### Files
There are a few files included in this repo.  The technical report ipython notebook includes code related to data collection, exploratory analysis, and model building.  It contains two primary models, one a random forest model that predicts annual energy production, and an elastic net model that predicts installation cost.

The data cleaning notebook contains info regarding processes for cleaning data collected from the OpenPV project, and dataset of over 1 million solar panel installations across the US.

The NREL query python file is a script that was run every 24 hours for 2 weeks on an AWS EC2 instance in order to gather solar radiation data for the 15,000 ZIP codes worth of data included in the OpenPV dataset.

### Questions
for any questions about the work here feel free to email me at evancasebaker@gmail.com.

solar-calculator-models

solar-calculator-models Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

solar-calculator-models

solar-calculator-models Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

Download README