AI for Capital Planning

The Government of Alberta is working with AltaML to enhance education infrastructure spending decisions.


The Province spends between $200M and $700M annually on new capital projects for schools. The average school utilization rate across the province is 71%, but the target rate is 85%.

A 1% increase in average provincial utilization represents a significant impact to students and school authorities and could lead to $19.6M in annual savings.

The Challenge

The Government of Alberta’s Education Capital Planning department approached AltaML with a challenge: help them use their data to improve how they allocate capital. Every year they receive hundreds of applications for new schools and they need a data-backed solution to help evaluate and prioritize these projects. If they can determine the optimal location and size of new schools, it will allow them to improve school utilization rates. However, this is a complex challenge as there are many factors that affect enrollment including location, programming, grade, and more.

A significant problem was predicting the enrollment rate of these proposed new schools. Lack of insight into expected enrollment can lead to underutilization or overutilization of that school, which can both be very costly. A 1% underutilization of schools across Alberta represents $19.6 million in annual waste. This is the cost of building a 300-student elementary school.

The department has access to a significant amount of data on students, and they need it presented in a user-friendly way, along with insights that allow them to make decisions on new school applications. Solution

A machine learning (ML) model was built that allows users to place a hypothetical new school, and predicts the enrollment pressure on that school and surrounding schools. This model is presented with a clean user interface that provides an efficient means to support the decision-making process, allowing the team to prioritize, analyze, ideate and collaborate.

Prioritize: The Capital Planning Manager can use the ML tool to identify where the school will be more efficiently utilized and more accurately assess the priority of applications.

Analyze: Using spatial analysis and ML, the Capital Planning Manager can stop using spreadsheets and, instead, review student distributions across a map. 

Ideate: The Capital Planning Manager can generate alternative scenarios, recommending appropriately sized facilities for each community. They can help avoid future low and high utilization by generating and analyzing more scenarios. They can more accurately predict appropriately-sized school core build and future additions of modular classrooms when enrollment justifies.

Collaborate: The Capital Planning Manager can use the tool in discussions with the school boards to review current usage, future demand and alternative scenarios to improve utilization as well as improve proposal review times.

Future Work

The GovLab team is working with the Education Capital Planning team to build additional features and get the tool ready for a multi-city pilot in the spring of 2023. In addition, we are exploring the possibility of bringing these benefits to other regions.

For more information, please see our brief video or contact us at