Skip to content

Implementation of the work presented in the AAAI 2019 paper: "Predicting Hurricane Trajectories Using a Recurrent Neural Network"

Notifications You must be signed in to change notification settings

sheilaalemany/hurricane-rnn

Repository files navigation

Predicting Hurricane Trajectories Using an RNN

In this project, we use a recurrent neural network to predict the trajectory paths of hurricanes. We use the Hurricane Database dataset from Unisys Weather. Here we preprocess the data, train a recurrent neural network on some of the samples, and compare our predictions against the real hurricanes. So far, we're using only the years 2000 to 2009 and is saved on our repo as '2000-2009-data.csv'.

Getting Started

The model was implemented using Python 3.6. The code was written and run in a Jupyter notebook. For local installation, make sure you have pip installed and run:

$ pip install notebook

Then to run, you would launch the notebook with:

$ jupyter notebook

Main Tools Necessary

Running

The project successfully trains a model, however graphing the trajectories (latitude and longitude points) needs work. That graphing is under construction.

About

Implementation of the work presented in the AAAI 2019 paper: "Predicting Hurricane Trajectories Using a Recurrent Neural Network"

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published