Tasks for teaching predictive modelling and APIs often require learners to use code-driven tools. Minimal research, however, exists about the design of tasks that support the introduction of high school students and teachers to these new statistical and computational methods. Using a design-based research approach, a web-based task was developed. The task was constructed using our design framework and implemented within a face-to-face professional development workshop involving six high school statistics teachers. The teachers were guided through the process of developing a prediction model using: an informal approach; visual prediction intervals; data about movie ratings from an API; and R code that ran in the browser. Our findings from this exploratory study indicate that the web-based task supported the development of new statistical and computational ideas related to predictive modelling and APIs.