How Data Scientists can help Government Agencies Effectively Respond to Natural Disasters

Case study of wildfire impact in San Diego County

The availability of information for disaster preparedness, recovery, and response is ever changing. Social media and other online platforms have freely available data that can help local and Federal response agencies to understand potential and real impacts from a natural disaster. Additionally, this data can be used to stage and deploy resources when they are needed to mitigate these impacts. Our team sought to create a replicable process for identifying businesses in a given geographic area, and labeling and displaying them according to the FEMA Lifelines protocol. We used Yelp and Google data, as well as a multitude of different mapping technologies. We established a repeatable process that can be used to map both potential risk based on historical information, and would allow FEMA and other agencies to also map current disasters to show in real-time what types of resources will be needed in the disaster area.

When a disaster strikes, the ability to quickly gain an understanding of the potential impacts to response efforts is vital to resource allocation. The United States sees more than 80 natural disasters every year, causing extensive property destruction and loss of life. The Federal Emergency Management Agency, or FEMA, works with first responders to support citizens in the preparedness for, response to, protection from, and recover from these natural disasters. FEMA is regularly faced with the challenge to manage the efficient and timely response to disasters, as well as anticipating and meeting the needs of local agencies and populations.

In recent years, the frequency and impact of wildfires in the United States has exponentially increased, with California being one of the main states impacted. In 2018, there were 8,054 fires in California that burned 1,823,152 acres, compared to in 2010, when 6,502 fires burned 108,742 acres. Additionally, there are significant financial and human losses as a result of these fires. In 2018 alone, there were 85 lives lost and a $400 billion impact to the state. San Diego and its surrounding communities in particular have been affected by some of the largest impacting fires in the last two decades. The Cedar fire in October 2003 claimed 273,246 acres, 2,820 structures, and 15 lives from San Diego County. Also in San Diego County, the Witch fire claimed 197,990 acres, 1,650 structures, and 2 lives in October 2007. Both the Cedar and Witch fires are among the top 10 most destructive fires in California’s recorded wildfire history.

The challenge that we set out to address was how we can help FEMA to identify businesses that may have been impacted in a disaster zone, and which Lifeline those businesses may align to. Utilizing this information, we aimed to create an interactive map for visualizing these Lifeline locations in relation to areas that have historically been affected by disasters and areas that are at risk to be affected in the future. We applied this technique to San Diego County, however the idea was to create a method and results that could be carried out for any other locations in the United States. The method and results were presented directly to New Light Technologies, one of FEMA’s contractors through General Assembly.

The tools we used in preparing this analysis and technical report include the following:

The following data sources are included in our analysis:

In addressing this problem for FEMA and New Light Technologies, we used the following approach:

1. Collection and Cleaning of data

2. Align the Lifelines

3. Mapping

4. Conclusions and Next Steps

Our original prompt suggested using Yelp API data, which we set out to do. We quickly learned, both from looking at the data we received and consulting previous GA students’ projects, that this data is not very robust and did not include enough information to solely rely on Yelp. We decided to also incorporate data from Google Places utilizing their API. In addition to the location data, we had to use definitions and descriptions of what the actual FEMA lifelines are from FEMA’s website. This formed the basis of the data we used for mapping.

In order to add context to the map, we collected data on historical disasters affecting the San Diego County area. We found specific perimeters of the areas in the county affected by some of the historically most damaging fires. We also found data from USDA that shows the current wildfire high risk zones, which was also added to our map.

Once our data was collected and aligned with a FEMA lifeline, we mapped those locations out along with the historical disaster-affected areas and the current high risk zones. We wanted to provide a map that was intuitive and provided the most value, so we tested a few different tools before deciding on a final option.

While this was easy to map, the image was not very clear and was not very interactive. It was also not possible to map the Lifeline businesses and then add layers of fire perimeters or current fire risk zones.

We could map the Lifeline data using a csv, however we could not clearly distinguish between Lifelines.

Ultimately, we decided ArcGIS provided the best usability, interactive capability, and visuals. We could apply many layers to this map and update those layers simply and efficiently. We included layers of all businesses color-coded by Lifeline, the perimeters of historical wildfire disasters, and current high risk fire zones. The next few images demonstrate some of ArcGIS’s capabilities.

For a video demonstration of the full functionality of the ArcGIS map we created, take a look at the following visual:

Using data from Google, Yelp, FEMA, and the USDA, we were able to create a map that shows the businesses in San Diego county and to which FEMA Lifeline those businesses align. Additionally, we were able to add layers to the map of the perimeters of areas that were previously affected by a wildfire disaster or are in a high risk area that could be affected in the future. This information can help FEMA decision-makers to mobilize and anticipate resource needs during a disaster.

The map is a tool that can be used as a starting point to determine evacuation plans, or estimate potential or actual impact from a disaster. Knowing locations and which lifeline they map to, can help disaster response teams more quickly help those locations or provide aid to those who need it during a disaster.

We recommend to New Light Technologies and FEMA to invest in public and private technology partnerships. With the valuable tools we found to complete this analysis, we were able to paint a picture of how businesses and access to Lifelines could be impacted by a disaster. With less barriers (scrape limits/delays, software costs), our process could be reproduced in minimal time for any location in the United States. Technology partnerships would greatly increase the efficiency of identifying and mapping a disaster to assess its impact.

Additionally, if a partnership was established with platforms like Yelp or Google, they could add new features to better help FEMA in their response. Adding visible fields on their sites that note if the business is in an area affected by a disaster and whether or not that business is open as a result of the disaster could streamline our method even further.

Additional technology investments would help FEMA and their challenges in responding to and addressing disasters. These investments could include expansions on the tools that we tested here (ArcGIS) or investment in new tools. The ArcGIS tool that was used to create our map was built through a free trial. For next steps, we recommend investing in this type of software to enhance this map and apply it to other areas affected by disasters. It has a software usage fee of $500 per year.

The following sources were imperative in preparing this project:

Please see the links to the rest of the team’s Medium articles (we cannot co-author stories on Medium):

Brenda Hali, John Kirby, Larry Curran

Related posts:

The physical benefits of eating healthy are constantly discussed, but to me it seems like the connection between food and mental health is often ignored. This the biggest reason I changed the way I eat, so here's some of what I know about the connection of food and mood...

What was the enigmatic Nikola Tesla like in his childhood? Or in his twenties? This is a concise, detailed biography of the lone genius in his formidable years.

Our hometown is our sincere friend. We leave it for our future life but it keeps calling us. Whenever we meet it, it opens arms for us and its hug touches our soul. It always keeps our dream saved…