How the Red Cross uses address data consolidation to save lives

Running a program like this on a national scale poses many challenges: Having enough volunteers, getting accurate data from national statistics, and using that data to correctly predict areas of high risk for preventive visits. Prior to the launch of the Home Fire Campaign, the Red Cross operated several boots-on-the-ground strategies around fire preparedness, including canvassing neighborhoods with information on testing smoke alarms and other fire safety tips. They also partnered locally with a few fire departments to help facilitate smoke detector installation events in some cities.

However, they saw a clear need to increase the reach of their fire prevention services. The agency was responding to more home fires (nearly 60,000 per year) than any other type of disaster—tornado, hurricane, etc. They wanted to identify where they could make the biggest impact and decided upon creating a more targeted home fire prevention program. But first, they needed to fill in some data gaps.

The U.S. Fire Administration collects data from fire departments around the country on an annual basis. It’s not a required activity in every location, but is strongly encouraged by the federal government, resulting in about a 70% report rate. The paperwork for reporting fire incidents is lengthy and passes through many different hands before getting compiled annually after which it’s made available to organizations like the Red Cross. Because the data is coming from multiple different fire departments and disparate sources, the address elements aren’t standardized, nor are they always valid.

Using Smarty US Address Verification, the Red Cross team was able to clean up, consolidate and standardize all of the mailing addresses in the collected range of sources. They uploaded addresses in bulk for verification and used the latitude/longitude information to geocode and create accurate maps. With an accurate database of address information in hand, the Red Cross partnered with DataKind, a non-profit data science group based out of the Washington, D.C. area to create a home fire risk map and predictive models.

First, they used large volumes of data on previous fire incidents and Census variables to estimate which areas were likely to have both high fire risk and a low likelihood of having working smoke alarms. They’re continuing to work with DataKind to build and refine their models to reliably predict home fire risk. Because Smarty US Address Verification cleanses and validates address data, including second-level addresses like building units, the Red Cross can drill down to tighter neighborhoods where the predictive models say the risk is the highest.

A key differentiator of the campaign is the American Red Cross’s ability to track how many lives have been saved as a direct result of the home fire campaign. Using the standardization tools from Smarty, they can compare datasets to make accurate address matches between homes they’ve visited as part of their prevention measures and homes on the national data reports. Then, teams of volunteers go to the physical locations and interview responding fire departments, the local Red Cross teams that responded, and the clients served after the fire to find out all the factors involved, including whether it was the Red Cross smoke alarms that alerted them and if they used the home escape plan created with the Red Cross.