Building Causal Data Stories

This course teaches students how to think about research – from data collection to data analysis to reporting results – as a narrative process. Rooted in basic statistics, the course provides students who have some numeric literacy a clear and direct path for advancing from reportable results to stories with impact. Students will especially learn how to construct numeric stories that demonstrate causality.

Learning outcomes: Enumerate and defend the core relationships between research design and story-building, as well as define a valid causal inference and possible fallacies. Generate hypotheses based on data, explain how the hypotheses could be tested, and show how to translate different experimental designs and data pools into stories

Narrative Visualization

This course provides key strategies for visualizing results and enhancing the visual communication of data stories, followed by a deeper dive into how to go from analysis of results to a compelling story enhanced with visuals.

Learning outcomes: Create specific data visualizations that carry forward casual narratives, such as stacked bar charts, time series, and multidimensional causal chains, using effective and recognized visual and charting building blocks. Become familiar with how to avoid common pitfalls in designing visual content with narrative goals. Identify story elements that map onto the five narrative techniques learned and demonstrate through written compositions the ability to use them effectively.

Data and AI Storytelling as and Ethical Choice

In this course, students learn how to create data stories that use ethical research and protect basic human rights: autonomy, fairness, trust, beneficence, and non-malevolence.

Learning outcomes: Define and use appropriate checks and reasoning processes for ensuring data transparency, validity, and ethical use. Identify the effect of a story by the amount of belief change it induces in an audience, determine the threshold of significance for various audiences, and use the amount and type of learning for each story to predict the degree to which it could be an actionable story or not.