This is data that can save the lives of citizens and of first responders. By being able to predict where and when a fire will happen, or knowing the factors which will impact response time, or even predicting a fire rescue team’s workload for the coming tour. These aren’t wishful, hypothetical scenarios — they are the very real and fundamental needs of first responders that AI and data can meet — highlighting why technology must be at the forefront of every fire department’s toolbox. This article has been written by Dr. Lori Moore-Merrell, President and CEO at International Public Safety Data Institute (IPSDI). IPSDI helps local public safety agencies gather, organize and translate data to improve how they evaluate risks, deploy resources, and respond to emergencies. International Public Safety Data Institute (IPSDI) is a part of Public Spend Forum’s GovMarket Growth Program, as well as Public Spend Forum and Shatter Fund’s Women-Led Tech Accelerator Cohort.
In 2006, Clive Humby, who is a British mathematician, coined the phrase “Data is the new oil.” Later, George Firican wrote that “Both oil and data can be transformed into different products. From oil, you can produce anything from gas and plastics to detergents, toiletries, dyes, and movie film. Data can be converted into information that fuels human and AI [artificial intelligence] decision-making processes, which in turn enable self-driving cars, improve a company’s efficiencies, develop speech-recognition software, find cures to diseases, and much more.” These two statements only begin to describe the importance and the power that data can mean to the fire service. Data—and the knowledge that it imparts—is, in fact, the lifeblood of the fire service for the future.
The ability for fire service leaders to explain their department’s value is essential to protecting or enhancing resources for emergency response, training and prevention. Fire/emergency services leaders have access to massive amounts of data.
Structured data typically are well organized and easily formatted in searchable databases and include incident information, such as incident numbers and response times. Unstructured data have no predefined format—thus, much more difficult to analyze—and include social media, dispatch radio recordings and traffic cameras. Data capture, procurement and preparation of both types of data are fundamental to assuring sound analytics and data visualization.
As fire and emergency services departments become increasingly data-driven, ensuring access to internal and external analysts and data scientists is essential. As data sources become increasingly nontraditional, departments must access trained researchers and data scientists who can handle multiple datasets. Data scientists are trained to use technology as well as scientific methods, analytical models and detailed algorithms to mine intelligence and insights from structured and unstructured data.
AI is the sophisticated statistical analysis of massive amounts of data. Most AI today is known as narrow AI, which functions from engineered scripts to mine datasets and generates results. One type of narrow AI is machine learning (ML). ML has great promise for the fire service when given a consistent data feed. For example, quality incident data coupled with time of day, geolocation, and community hazard/risk data can be used to “train” ML models. ML then can be asked to draw conclusions based on observed examples of tasks. For instance, apparatus move-ups often are necessary during busy times with heavy apparatus deployment. ML can assess data from various data feeds to determine where the remaining (unassigned) apparatus should be relocated.
ML involves searching data for trends, patterns and anomalies that might not be obvious to a human observer. In the case of an emergency response system, an ML algorithm would learn to send proactive alerts when apparatus deployment thresholds are exceeded.
It also is possible to train ML response models with unstructured data that tend to be qualitative in nature. Data types might include social media, radio communications and video from body cameras. With this information and structured data, machine-learning algorithms for response-force model classification can be created. Text-mining and natural language processing can be used to extract intelligence from radio communication, which would contribute to more accurate apparatus move-up models.
A data-driven future
Fire and emergency services departments should prepare for increasing data integration into everyday activities. Leaders must gain greater data acuity for responsible decision-making. Fire chiefs must ensure that they allocate financial resources for personnel and technological capability for data capture, management, protection, governance, analysis and intelligence translation. Firefighters must become increasingly data literate, to understand the value of accurate data entry and report writing.
Because the importance of using data no longer is a question, the major challenge that departments face is how to process more data faster—for preparedness, prevention, operational insights, and firefighter safety and well-being.