Types of Projects

Cities participating in the City Data Alliance undertake resident-facing data projects. Each project centers on a clearly defined solution that can be scoped, tested, and adapted in real city operating conditions. The typology below describes common types of resident-facing problems in which data and AI can help cities make better decisions. Many problems span more than one category. These types offer a starting point for scoping and alignment, while allowing flexibility in how cities shape their solutions.

Who to focus on

Targeting & prioritization

These projects help cities identify which residents, locations, or cases to prioritize when needs exceed capacity, enabling earlier and more effective intervention.

Decision: Who should receive attention or resources first

Example projects

  • Identifying households most at risk of eviction to prioritize prevention services
  • Flagging properties most likely to present safety risks for proactive review
  • Targeting outreach to residents likely eligible but not enrolled in benefits

Where to invest limited resources

Allocation & optimization

These projects help cities make clearer tradeoffs about how to deploy constrained resources to maximize impact across neighborhoods, services, or populations.

Decision: How limited staff, time, or funding should be deployed

Example projects

  • Optimizing inspection routes or schedules to reduce response times
  • Allocating outreach staff across neighborhoods based on projected demand
  • Balancing caseloads across teams to improve service quality

How staff make day-to-day decisions

Frontline & supervisory support

These projects focus on improving real-time or recurring decisions made by frontline staff or managers by providing timely, relevant, and actionable information.

Decision: What staff or supervisors do in the moment

Example projects

  • Supporting caseworkers in prioritizing daily tasks
  • Helping supervisors identify bottlenecks or backlogs early
  • Improving coordination across teams working on shared cases

How residents move through services

Service pathways & experience

These projects help reduce friction, confusion, or drop-off by improving how residents experience city services across touchpoints and departments.

Decision: How services are designed and delivered to residents

Example projects

  • Redesigning intake or referral processes to reduce delays
  • Identifying where residents disengage from programs and why
  • Improving access for multilingual or underserved communities

Which path to pursue

Policy decision-making

These projects support leaders in weighing alternatives, modeling potential outcomes, and testing approaches before making or adjusting major commitments.

Decision: Which program, policy, or model to adopt

Example projects

  • Modeling how different shelter placements would affect system flow
  • Testing a proposed eligibility change with a small segment before full rollout
  • Comparing alternative deployment strategies using scenario simulations




A note on advanced analytics and AI

The City Data Alliance encourages cities to thoughtfully explore how advanced analytics and AI can strengthen decision-making and improve resident outcomes. This may include approaches that surface patterns, risks, or opportunities from large or complex datasets. The City Data Alliance supports these efforts when they are grounded in real operational decisions, paired with human expertise, and used responsibly in live city contexts.

What is typically not a good fit

The City Data Alliance is not designed to support projects that are primarily strategic, exploratory, or technology-driven without a clear path to action. This includes standalone data strategies or policy development efforts without a specific resident-facing solution, dashboards or reporting that are not tied to concrete decisions, technology procurement or vendor selection without the opportunity for testing and iteration, academic research without operational ownership, or citywide AI strategies that are not grounded in a near-term, testable problem.