Is Domo just woeful at even the most basic analysis tasks or am I missing something?

user013672

I have what I think should be a very simple and common data analysis task. And yet I always have to resort to pulling the data down, manipulating it myself locally, then pushing/syncing only the absolute final numbers into Domo that slot perfectly into a bar chart.

 

An example:

I have a few million rows of data. That data has timestamps, categories, and usernames in their respective columns.

I want to get a count of how many users used each category in the last 7 days. How would I go about doing that level of wrangling within the Domo platform? And then a follow-up would be, what percentage of users (given each user belongs to a company) for each company used X category.

 

Because my current solution, which is generating a row of "card information" per row per time period per category, explodes dramatically with number of features. Especially because it would have to be repeated with small modifications for different cards/visualisations. What I currently have:

Row 1: Category A1, Last 7 days, Used 10 times
Row 2: Category A1, Last 14 days, Used 18 times
...
Row 9999: Category Z9, Last 3 Months, Used 400 times

My current method, while functional, essentially avoids using Domo as much as possible and leads me to believe that I've not fully understood the capabilities of Domo.

 

I look forward to reading your replies. Thank you!

tadashii

Hi - this can be achieved by doing a lil data modeling. Assume below is the sample data and supposed today is 28 Oct 2019. The red block = last 3 days, green block = last 4 days, the orange circle is just to illustrate that unique user for category c in the last 3 days.

 

 

 

Use magic ETL to join the sample data to the "Dim last x Days" dataset. ( You can use SQL to auto-populate the date against the "last x Days " label )

 

 

The card, just drag in what you need. 

 

You can use beast mode for the 2nd question.

 

Hope that gives you some idea.