1

I have a dataset consisting of unique trips (IDs), users_id, and bikes_id and time-stamp trips_id, users_id, and bikes_id are dimensions:

  1. A trip_id is unique
  2. A user can have multiple bikes_id and trips_id
  3. A bike can have multiple rides and multiple users

I would like to show how many rides there are (on average) per user, how many bikes are used (on average) by user and how many rides there are per bike. Is there a calculated field or LoD that can do this calculations? So I can show per week, month or another time period?

AnilGoyal
  • 25,297
  • 4
  • 27
  • 45
sanchezjAI
  • 91
  • 1
  • 9
  • Could you please include a minimal amount of sample data as input and desired output in respect of that data? – AnilGoyal Dec 24 '20 at 06:10

1 Answers1

3

I would like to show how many rides there are (on average) per user

use this CF (average for whole data)

{AVG({FIXED [USER] : COUNTD(TRIP_ID)})}

how many bikes are used (on average) by user

use this

{AVG({FIXED [user]: countd(bike_id)})}

how many rides there are per bike

use this

{AVG({FIXED [bike_id]: countd(trip_id)})}

NOTE these calcs will give you average for whole dataset

AnilGoyal
  • 25,297
  • 4
  • 27
  • 45
  • I have already tried that, but my idea is to show the average per week, month or other time dimension. so considering the "datetime-stamp-dimension". – sanchezjAI Dec 25 '20 at 02:38
  • What does average per week mean? I am nit able to understand! Can you please include some sample data and output required? – AnilGoyal Dec 25 '20 at 05:27