Level 3 - Question 53

Question

Return an activity histogram that shows users by grouping them together based on the number of active days they have had in the previous month.

Additional Context

Build a query in Explore using Bar view. The query should use the actor property built in question 52 that returns a count of unique user actors, filtered to all user actors, and is split by the actor property you created in question 52. In measure 1 be sure to set the "as a fraction of.." to all. Set the time window starting from 1 month ago and ending now. When the query returns, use chart options to format measure 1 as a percent, and at the bottom select "display value of each bar at the top".

Take Note:

This query is doing a very similar thing to the last few questions we worked through. But instead we’re counting the number of active days that a user has had in a given month so we can return a histogram of users based on how many days they have had in a specific month, versus counting binary criteria of whether they were or were not active over a several month-long period. Here we’re able to easily see what the average or most common number of days your users are active. In the results, we can see that every user who has had 7 days has been grouped into the 7-day bucket, and every user who has had 6 days has been grouped into the 6-day bucket, and so on. This is because the actor property we created counts the number of days that each user had in that time period and returns a value for them. In this query we’re simply saying count users and split them up by the bucket they fall into in actor property 52, show it as a bar and as a fraction of all users.

Solution

Note: The numbers and result do not need to be an exact match as the data changes over time. The query construction is what is most important. See images below.

*Click to enlarge*