In May 2015, journalist Maggie Lee posted the following to an email list of data journalists: “So, I’m a state gov’t reporter learning CAR, hooray. Python and D3 and whatnot etc. But Lord, what should I make?" Lee’s question, of what to write about, is a common one for digital journalists armed with tech skills but seeking an appropriate story on which to unleash them. Some journalists, asking themselves the same question, will undoubtedly answer, “I want to do an investigative story that uses government data to uncover potential fraud.” Given the growing interest in computational analysis inside digital journalism and the resulting increase in questions such as Lee’s, I offer a method for constructing a software system that will assist in developing a suite of story ideas on a single theme on a public affairs beat. The system, which I call a Story Discovery Engine, is a reporting tool. Like any specialized tool, its purpose is to help its user with a specific task, but it will also be useful for other as-yet-unimagined purposes. Creating a Story Discovery Engine and deploying it on a beat can help reporters to come up with a variety of story ideas based on data analysis.