Machine Learning for Proactive Cities

Technology is impacting our cities in a variety of ways; one of the most exciting impacts relates to the rise of Machine Learning. Machine Learning, as a computer science area of study, applies algorithms to various datasets to discover patterns of interest.  It is a procedure through which knowledge is acquired by establishing algorithms constructed to “learn” about the related needs of various issues, bringing them into a broader, structured pattern.  From this learning, the potential to predict what might happen in the future and respond proactively and preemptively is derived.


A few established applications include:

1. Vehicle Collision Incident Reduction – By using Machine Learning to parse together data about vehicle collisions, such as police incident reports and 911 calls, with data about the surrounding context, such as large public events, precipitation, and damaged infrastructure, a picture about contributing factors can be painted. Trend analysis can highlight interconnected factors across many vehicle collisions, clarifying which factors contribute the most harm. Armed with these insights, cities can respond with effective engineering and design solutions to proactively reduce vehicle collisions.

2. Crime Reduction – Crimes are often committed in a series over time by the same perpetrator and have common characteristics related to the perpetrator such as weapon type or brand, interaction with victim, spoken phrases, or targeted theft items. These characteristics can be tracked across multiple past crime incidents via Machine Learning, Natural Language Processing (NLP) more specifically, to pinpoint potential correlation, providing investigators with insights into which crimes may be associated with the same perpetrator. By pinpointing this correlation earlier in a crime sequence than would be possible without Machine Learning, future crimes can possibly be prevented altogether.   It also enables a police force’s time to be used more efficiently, increasing the quality of outcomes with the same tax dollars.


As Machine Learning transforms how cities proactively understand interconnectivity and strategize tactics in response, it provides insight into how various issues are addressed. In the Machine Learning space, it is important to see how inroads are being made on certain issues (e.g., transport, public safety, and others) in case they can be applied in other cities. In addition, parallels can be drawn to inform other issues (e.g., such as environmental sustainability and affordable housing) as the logic and structure behind Machine Learning is further developed.

ComplexCities is in the process of establishing a professional network and pinpointing resources to share for this project.  If you are interested, please contact us.