Introduction to Machine Learning for Crime Analysis

Introduction to Machine Learning for Crime Analysis is a 12-week online course that introduces machine learning in a practical, applied way. Students will learn how to use machine learning algorithms to find patterns in crime data, make analytical decisions from those patterns, and choose models that fit the problem at hand. They will then apply those models to predict, classify, infer, and explain crime events in real analytical settings.

The course emphasizes the mathematical ideas behind data processing and statistical inference, but it is not a math class. Modern packages handle most of the computation; our focus is on what the models are doing, why the mathematics matters, and how this affects interpretation, bias, and defensible conclusions. We also look at how machine learning overlaps with traditional statistical analysis, where it differs, and why those differences matter in operational work.

Topics covered include statistical inference, regression and classification techniques, model validation methods, decision trees and ensemble models, clustering and dimensionality reduction, and the practical use of machine learning to analyze and interpret crime data.

Instructor: Salena Torres Ashton

At a Glance

Upcoming Sessions

Select the session that works best for your schedule. Once a session has passed or is already in progress, registration is disabled.

Note: availability labels are updated manually as seats fill. If a session does not appear on the registration form, it should be treated as sold out.

Class Format

This is an advanced course for crime analysts who have prior experience with Python (or a similar language), a working foundation in statistics, and experience with crime analysis. Students should be able to:

• Use core data structures (lists, dictionaries, sets, or equivalents) in Python, R, or a similar language
• Write functions and pass arguments comfortably
• Apply basic statistical concepts, including population vs. sample means, variance, standard deviation, and sources of error
• Work with foundational algebra, including expressing numerical patterns as formulas and reasoning about rates of change
• Work with the same dataset across the full 12-week course

Students should expect a time commitment of 3-5 hours per week to gain a passing grade. The course is graded as pass/fail. This class follows a weekly-structured online format, meaning the class advances through each topic together week by week.

Students can log in whenever it fits their schedule, and there are no live sessions. Participation throughout the course is required for a passing grade. As a rule, bulk submissions during the last week of class are not accepted unless prior authorization has been granted by the instructor.

Invitations to the learning platform are sent the week before class begins.

Course Outline

This course is organized into 12 weekly modules that build cumulatively, with each week extending previous material. Weekly projects are progressive and designed to mirror real analytical workflows.

Ready to Enroll?

Choose an upcoming session and complete your registration. If a session fills, students may contact [email protected] to ask about availability or future offerings. See the Training Policies page for information on transfers, participation rules, and refunds.