When the app displays how much an event is promoting or detracting from an intention, currently this is based on a Pearson correlation. However correlation doesn't always indicate causation. Improve the analysis by using a causal inference model like Python's DoWhy.