Devpost
Participate in our public hackathons
Devpost for Teams
Access your company's private hackathons
Grow your developer ecosystem and promote your platform
Drive innovation, collaboration, and retention within your organization
By use case
Blog
Insights into hackathon planning and participation
Customer stories
Inspiration from peers and other industry leaders
Planning guides
Best practices for planning online and in-person hackathons
Webinars & events
Upcoming events and on-demand recordings
Help desk
Common questions and support documentation
Retrofit Radar predicts how much energy a building uses, then flags the buildings that use far more than comparable buildings in the same climate, so an organisation knows which ones to retrofit first
This MATLAB script predicts building energy use (site EUI) from building and climate data using gradient-boosted trees, checks its accuracy on held-out data, and saves test predictions to a CSV file.
Leak-free, physics-informed MATLAB model predicting building energy use (RMSE 37.6, 34% better than baseline) that turns predictions into a retrofit shortlist of 1,123 high-waste buildings.
Model to predict energy consumption
A building’s past predicts its energy future, built entirely in MATLAB.
The task is to predict Site EUI, the energy a building uses per square foot per year) for 9,705 buildings in a future year. The inputs are building characteristics and the local weather.
Retrofit Radar predicts how much energy a building should use based on its type, size, age, and climate, then flags buildings that use far more than similar ones.
We compared linear regression, decision trees and boosted trees in MATLAB to predict building energy use. Boosted trees were best, cutting error by about 29% versus guessing the average.
We predicted building energy use from 75k records. Bagged trees cut error 24%, and the data shows where to retrofit first: buildings rated under 10 on Energy Star use 2-6x more than top-rated peers.
nil
1 – 10 of 10