Smoke Signals: Open Data and Analytics for Preventing Fire Deaths

Smoke Signals: Open Data and Analytics for Preventing Fire Deaths
By

Today we are launching Smoke Signals, an open source civic analytics tool that helps local communities determine which city blocks are at the highest risk of not having a smoke alarm.

25,000 people are killed or injured in 1 million fires across the United States each year. With over 130 million housing units across the country, 4.5 million of them do not have smoke detectors, placing their inhabitants at substantial risk. Driving this number down is the single most important factor for saving lives put at risk by fire.

Organizations like the Red Cross are investing a lot of resources to buy and install smoke alarms in people’s homes. But a big challenge remains: in a city of millions, what doors should you knock on first when conducting an outreach effort?

We began working on the problem of targeting the blocks at highest risk of not having a smoke alarm with the City of New Orleans last spring. (You can read about this work here.) Over the past few months, with collaboration from the Red Cross and DataKind, we’ve built out a generalized model and a set of tools to offer the same analytics potential to 178 American cities, all in a way that is simple to use and sensitive to how on-the-ground operations are organized.

We believe that Smoke Signals is more a collection of tools and collaborations than it is a slick piece of software that can somehow act as a panacea to the problem of fire fatalities. Core to its purpose and mission are a set of commitments:

  • an ongoing collaboration with the Red Cross wherein our smoke alarm work informs their on-the-ground outreach
  • a collaboration with DataKind to continue applying volunteer work to the improvement of the underlying models and data that drive the risk analysis
  • a working relationship with major American cities to help integrate our prediction models into their outreach programs
  • a downloadable CSV for 178 American municipalities that associate city streets to risk scores
  • an interactive map for an immediate bird's eye assessment of at-risk city blocks
  • an API endpoint to which users can upload a CSV of local fire incidents in order to improve scores for their area

We believe this is an important contribution to public safety and the better delivery of government services. However, we also consider it a work in progress, a demonstration of how civic analytic solutions can be shared and generalized across the country. We are open sourcing all of the components that went into it and invite anyone with an interest in making it better to get involved.

It's not hard to imagine a dozen variations of Smoke Signals using public data. Quantifying the at-risk cross-section of people with disabilities and flood zones? Doable with the American Housing Survey (AHS) data that underpins our smoke risk models, and FEMA's national hazards geodata. Predicting which blocks in a city are most at need for roof repairs? Property tax assessments and the AHS, yet again. We believe that there are countless untapped sources of data and insight that can be marshalled to improve the lives of citizens. However, it’s impossible for any one organization to do more than scratch the surface of what’s possible. At Enigma, we’re always on the lookout for partners and collaborators in this mission. So if you’d like to get involved, please get in touch.