Hurricane Watch

• 

• 

  Changing sea temperatures over time (average of data points in the N Atlantic). A 0.5 degree increase in temperature is observed.

• 

  Maximum hurricane wind speed over time. The most catastrophic hurricanes are increasing in intensity.

Inspiration

We believe that information is the most important resource in creating a more sustainable world. Hurricane Watch uses years of existing datasets to aid in disaster response, and to highlight the importance of climate action.

Recent years mark some of the most destructive hurricanes ever recorded in the Atlantic, with the “ultra-intense” Hurricane Dorian devastating the Bahamas before hitting the U.S. and Canada, causing $4.6 billion in damages and breaking records as the second-most powerful Atlantic storm recorded. Hurricane damage continues to exceed forecasts, while policymakers continue to make cuts to funding disaster relief agencies. We wanted to investigate the impacts of anthropogenic warming on hurricane formation in order to spur climate action and further governance of the global commons.

What it does

We used 40 years worth of sea surface temperature, pressure, and wind shear in order to predict the frequency, duration, and strength of hurricanes. Graphs highlighting the relationship between different factors are presented in an easily-accessible way in order to inform policymaking. Our model highlights the most important features that affect hurricane formation and highlight the most at-risk regions on the east coast.

We collected data from four datasets (IBTracs, ERSST, Blended Sea Winds, NCAR Gridded Pressure) obtained from the National Oceanic and Atmospheric Administration and the National Center for Atmospheric Research.

How I built it

We used several Python libraries to analyze the thousands of data points mapped by coordinates and time. Data was plotted using matplot.lib in a Jupyter notebook, and we performed an xgBoost in our predictive model.

Challenges I ran into

Initially, it was incredibly challenging to process and visualize a large amount of data, but over time we were able to process the data so that it can be easily analyzed.

Accomplishments that I'm proud of

We were incredibly proud of making sense of the huge amount of data made available to us. Finally discovering the correlation between our different variables was an incredible moment, and we were very happy with the results.

We were excited to have been selected as a finalist, and for the opportunity to present our work to an audience.

What I learned

This was one of our first projects studying data analysis, so we were introduced to the process of researching, accessing, and analyzing datasets.

What's next for Hurricane Watch

Our core product for this hack was the presentation of the data, and we would like to improve the user interface and interactivity of this in the future.

Built With

• matplot
• python