-
-
Choosing location and time interval for prediction
-
Interactive map for shelters
About the Project
Our project aims to revolutionize the way we predict and prepare for power outages, ultimately minimizing their impact on communities. Inspired by the need for more effective outage management, we set out to harness the power of Machine Learning (ML) to predict outages based on weather patterns.
What Inspired Us
The idea stemmed from recognizing the importance of understanding outage causes and patterns, as electricity is a critical resource. We were intrigued by the potential of ML to enhance outage prediction accuracy and provide valuable insights for mitigation strategies.
What We Learned
Throughout the project, we learned the intricacies of working with large datasets, especially in the context of weather and power outage data. We gained a deep understanding of ML techniques for time-series forecasting and feature selection, crucial for accurate predictions.
How We Built It
Our project involved gathering and cleaning 15 years of power outage data from US cities, along with daily weather forecast data. We built a user-friendly website with interactive tools to predict outage probabilities based on weather conditions. Our tech stack included HTML, CSS, JavaScript for the frontend, Flask for the backend, and Python for ML modeling.
Challenges We Faced
One of our main challenges was finding reliable and relevant datasets, as well as cleaning and normalizing the data to remove noise. We also faced challenges in model selection and tuning to achieve the desired prediction accuracy. However, overcoming these challenges has been immensely rewarding, and we are excited to continue refining our model and scaling it to Canadian datasets.
Our project has the potential to significantly impact how communities prepare for and respond to power outages. By providing accurate predictions and actionable insights, we hope to contribute to a more resilient and prepared society.
Log in or sign up for Devpost to join the conversation.