Business opportunity

The City of Stockton serves more than 320,000 citizens in California’s Central Valley. The City had collected years of stored data they wanted to leverage using AI to improve citizen services and satisfaction. Its data management team determined that Public Works data was ideal for delivering an initial solution. To bring its vision to life and create an app with a smart dashboard for citizens, the team partnered with Kyndryl to help operationalize its data.

To create a Minimum Viable Product (MVP) to address the first identified use case, the joint team chose to focus on creating a machine-learning model for forecasting the time it would take to repair reported potholes, following initial citizen complaints.

For city residents and workers who travel the same route frequently, it’s frustrating to navigate around potholes and not have visibility into repair schedules. The City saw this MVP as an opportunity to ease these frustrations and generate taxpayer value while also creating efficiencies for roadwork crews.

Technical challenge

The use case was jointly identified by the City and Kyndryl because through discussions, Kyndryl learned that Stockton had gathered years of data on road damage and repairs, including GIS location data, construction dates, pavement materials and dimensions.

The City also had historical data on street conditions, workorder data, including repair start and completion dates. They also had access to industry standard information such as the pavement condition index and traffic index. This data was effectively invisible to citizens and underutilized by the City itself: there were no forecasting capabilities in place and existing web-based guidance was static in nature.

The vision was to create a dynamic, publicly accessible dashboard app, that would aggregate the selected public works data sets and use AI to continuously predict repair timelines for newly reported potholes. This would require significant data cleansing, the development and training of AI and machine learning models, and the creation, testing and validation of automation pipelines.

 

Our solution

Together, the City of Stockton and Kyndryl created a citizen-focused online dashboard that provides accurate forecasts for pothole repairs and helps road workers plan more efficiently.

Kyndryl Consult helped define a minimum viable product (MVP), secured funding from alliance partner, Microsoft, and co-created the solution in just 30 days. This included building:

• A machine learning regression model to predict pavement conditions
• A binary data classifier to determine 1-day job probability
• A forecast model to estimate repair timelines
• Pipelines in Azure Databricks providing daily updates to citizen-facing dashboards on Microsoft Power BI

The City plans to extend the dashboard to include information on traffic and public safety.


 

What progress looks like

Kyndryl Consult’s technical leadership and cloud expertise helped the City of Stockton create its first-ever public AI tool in under 30 days. Delivering real-time engagement and visibility into estimated repair times, the Kyndryl solution helps the City provide value to citizens and increases the credibility of the administration.

•  Launched first public AI tool in city history
• Gives citizens useful information through a user-friendly dashboard
• Provides city workers with more efficient planning insights with which to speed time to repair
• Leverages the City’s existing Open Data Portal for cost-effectiveness
• Boosts confidence in the City’s leadership
• Established framework for future smart city data products

See more examples of how Kyndryl is helping governments modernize at kyndryl.com/us/en/industries/government.