This paper discusses a growing tool of interest for cartel detection: examining market data for evidence of collusion, or what is referred to as cartel screening. Screening identifies collusive patterns in firm conduct such as prices and bids. The first half of the paper describes what to look for in the data, more specifically it features collusive markers, structural breaks, and anomalies. A collusive marker is a pattern in the data more consistent with collusion than competition. A structural break is an abrupt change in the data-generating process that could be due to cartel birth, death, or disruption. An anomaly is a pattern in the data that is inexplicable or inconsistent with competition. The second half of the paper focuses on the recent use of machine learning algorithms to develop more effective screens by extracting the most informative patterns from the data, which then instruct us what to look for in the data. With access to a data set comprising episodes of collusion and competition, supervised learning can identify patterns indicative of collusion. Proof of concept is shown based on work of the Swiss Competition Commission using data from construction cartels in Switzerland. Guidance is provided for other competition authorities to deploy machine learning algorithms, including deep learning, to make cartel screening more effective.