Running requirement: Linux, Python, Keras, TensorFlow, qt4 for GUI
Important Note: One line was changed, which will potentially help you run the program depending on your environment, check github for updates if it doesn't work on your computer.
Climate change is an important aspect of our future that needs to be in the mind of every American. However, it is often hard to relate specific realities of everyday life to climate change. This is why we made a tool to allow people to gain a more individualized understanding of the impacts of climate change.
This program allows user input of a city and a date in the future, and using deep learned neural networks, we make predictions about this date and provide information about what the future would be like.
Starting with NOAA datasets, we trained custom made networks for each location. We then made a graphic user interface, to improve usability for everyday Americans. We then linked the two programs for out final product.
A large part of this project was dedicated to data analysis. We had to find good climate data, then process it by hand and read it in to a program.
Simply building an entire new program from the group up was new for us, and it is an accomplishment.
(NOT WORKING WILL UPDATE) Having never made a GUI before, that was a new part of C++ that we learned.
Functionality will be greatly increased. The number of cities trained was severely limited by training time as was data visualization, but this has potential to be a powerful tool. What we want to do next, is to use a CNN and Landsat data to generate specific images of the future to further increase the users sense of importance of this issue
Log in or sign up for Devpost to join the conversation.