National Police of Colombia Uses AI-powered System to Improve Public Safety and Prevent Crime with Nuvu on AWS

With a population of nearly 53 million in a country of more than 440,000 square miles, the NPC must focus its resources to reduce crime and protect its citizens. To help it understand patterns of crime and determine where to best allocate officers, it wanted to better use the data it collected during its daily operations.

The NPC wanted to find a way to improve the analysis of crime data. The organization was conducting its own analysis using data from calls to its 123 emergency number, with the data entered into spreadsheets for processing. The results took about 2 months to complete for a single city. The NPC wanted to use its data to create a faster, proactive way to improve policing.

“We have a number of robust IT systems that didn’t integrate or talk together,” says Lieutenant Colonel Oscar Andres Cardenas Peña, chief information officer of the National Police of Colombia. “This made it difficult to pull that data together to quickly get a big picture of the policing needs of the community. That’s why we looked to using AI to help in forecasting criminal activity.”

The NPC chose to work with AWS Partner Nuvu, based in Bogota, an agency that focuses on digital transformation and specializes in serving the public sector. Nuvu works with TD Synnex, which provides the tools, resources, and the people to simplify the delivery of cloud solutions to IT solution providers to develop a proof of concept. A pilot project was launched in the city of Villavicencio that lasted for approximately 3 months. The first 2 months were primarily spent cleaning the data to be ingested. After the pilot proved the viability of the program, the project was opened to public bids for a full-scale initiative—a process that took about 6 months. Nuvu won this bidding process. In 2020, Nuvu started building the model at scale to cover Colombia’s largest cities comprising more than 40 million people.

“We have a history of providing AI solutions to the public sector,” says Andrés Barrantes Bernal, chief executive officer and co-founder of Nuvu. “The challenge organizations face is how to quickly and efficiently make use of data. That’s the benefit of AI. What used to take 15 days to process is now done instantly.”

The NPC and Nuvu worked together with TD Synnex to use its tools and platform to build a series of modules on AWS. The crime prediction module uses advanced algorithms and machine learning (ML) techniques to analyze various data sources, such as emergency hotline calls and crime records, to create heat maps that highlight high-risk areas and times for potential crimes. These predictions enable the NPC to proactively allocate resources, implement preventive measures, and take targeted actions to deter criminal activities.

There is also an optimal route module that helps reduce response times, and ensures the effective utilization of available personnel. A calls classification module uses natural language processing (NLP) techniques to transcribe, classify, and automatically prioritize emergency calls. The control panels module provides important metrics and insights regarding the solution's performance and adoption within the police department. It allows for monitoring key performance indicators, such as response times, route compliance, and overall solution effectiveness. This helps in evaluating the impact of the solution, making data-driven decisions, and continuously improving the system's performance.

“All these modules work together to provide quick, accurate access to the information the police need,” says Barrantes. “The use of AWS services helped us create, in collaboration with TD Synnex, an integrated policing system that can scale to help protect Colombia’s citizens.” The AWS services used include Amazon QuickSight, which is the tool in charge of AI and data science analytics models. Amazon Comprehend is used for NLP and Amazon SageMaker is used to build and deploy ML models. Amazon Kinesis allows the system to ingest data in near-real time, such, audio, application logs, website clickstreams, and IoT telemetry data, for ML, analytics, and other applications.

Nuvu’s ability to deliver a customized solution has played a crucial role in transforming the NPC into a proactive service. This customization and customer-centric approach has been instrumental in achieving exceptional results and meeting the unique needs of the organization. By using predictive insights, optimizing resource allocation, implementing targeted interventions, and ensuring prompt response to incidents, this shift not only enhances public safety but also fosters a sense of security and trust within the community by actively working to prevent crime and protect citizens. “The results of using AI to understand our data has resulted in a clear improvement,” says Lieutenant Colonel Cardenas, chief information officer (CIO) at the NPC. “Police response time has been reduced by 49 percent and we have improved the accuracy of our predictions by 89 percent, helping us more effectively deploy officers. At the same time, we have seen a 31 percent reduction in operational costs. AWS and Nuvu have helped us improve policing.”