For each individual user, it would be helpful to evaluate how their activity level compares to the rest of the cohort. We have a lot of raw data here, and it would be beneficial for us to run more advanced statistical analysis of the entire dataset, vs purely calculating individual metrics. E.g. if we could present percentile benchmarks across several variables, that would be a large amount of additional context when reviewing:

• of messages sent

• amount of code sent
  • (language used? count of top 5 represented?)
• number of reactions received
• longest streak
  • (some kind of consistency score?)
• ...??