Vulnerability in Infrastructure (To Seismic Activity)

Inspiration

The 7.8-magnitude earthquake in Nepal affected 3 million individuals and damaged 762,106 buildings. Analyzing the data gathered after this event can help us learn how better prepare for future earthquakes. Strengthening infrastructure can significantly enhance its resilience against seismic forces, and in turn, save lives.

We live in California, a seismically active state. So, we're especially interested in this topic.

What it does

This project uses data visualization to analyze the correlation between different building factors and the collapse rate in the 11 affected distracts of Nepal.

How we built it

This project is built using Python on Jupyter Notebooks. Our group mainly utilized data manipulation libraries such as Pandas, Numpy; data visualization libraries such as seaborn, matplotlib, and Altair; machine learning library such as scikit-learn.

Challenges we ran into

Originally, we planned to use the Reverse GeoCoder API by Melissa to determine exactly how far the buildings were from the epicenter. However, we ran out of time. We hope to develop this feature in a future model.

Furthermore, half of the team was inexperienced with data analysis. So, the learning curve slowed down the development of the project.

Accomplishments that we're proud of

While there was a steep learning curve for the beginners, we collaborated and produced informational data visualizations. We are proud that our work is meaningful and could potentially provide solutions for strengthening infrastructure.

While we don't often get to practice our data wrangling and machine learning skills, we're proud that we were able to practice and produce working models.

What we learned

As a team, we learned :

The basics of a data project (the steps, methods, etc.), as well as how to utilize the pandas library in Python.
How to present a formal data project and defend our findings.
Find connections between certain data points and make informative visuals.

What's next for Vulnerability in Infrastructure (To Seismic Activity)

As mentioned above, a future goal is to incorporate the Reverse GeoCoder API by Melissa. We also hope to improve the accuracy of our models from 75% to 90%.