Eazy BI - Epic size scatter plot and percentiles

[Nick Brown]

Hi guys,

Another question using Eazy BI :) 

I'm trying to create a scatter plot that visualises the number of child issues (story, task, etc.) within an Epic. X axis should have the Resolution date, Y axis should be the number of child items.

What I also want is to have a number/gauge visual above this that details for 25th/50th/85th/95th percentiles how 'big' an Epic is.

To give an example, this is the kind of thing I'm after:

But I'm not sure how to do this?

[Martins Vanags -eazyBI- · Accepted solution]

@Nick Brown 

You could create two separate reports:

One is scatter chart with two columns. First start by table report with a calculated measure "Epic resolution date" (formula below) and predefined measure"Issues created" for the columns and "Epic Link" dimension in rows. Then switch to scatter chart

Epic resolution date formula (with date formatting): 

[Issue].[Issue].Getmemberbykey(
[Epic Link].CurrentMember.key
).get('Resolved at')

The second report can be a gauge chart with the "values only" option enabled. This one would have 3 calculated measures where one of them (85th percentile) is calculated with the following formula. (I hope you can figure the other two by yourself):

Percentile(Filter(
Descendants([Epic Link].CurrentMember, [Epic Link].[Epic]),
[Measures].[Issues created]>0
),
[Measures].[Issues created],
85
)

And you can use The "project" dimension both in rows and page filter for the second report.

Then add both of the reports on the eazyBI dashboard with a common page filter

Martins / eazyBI team

[Nick Brown]

Thank you - that looks to work for me (I actually just used the 'Issue resolution date' measure), only question I have is around 'Issues created' as a measure. Is that right for showing child items?

Only for this particular Epic, there's differences in number created and number resolved:

When I look the Epic up in Jira, there are actually 11 child items...

[Martins Vanags -eazyBI-]

@Nick Brown 

Both measures "Issues created" and "Issues resolved" would also include epic itself if you use them with epics from "Issue.Epic" 

That is why I recommended earlier using the "Epic Link" dimension to not include epic itself.

But to answer the difference between both measures. each of two numerical measures in your report is independent.

If you filter the report by time, eazyBI would count how many issues were (tied to Epic) created (regardless of resolution details) in the filtered period, and how many were resolved. For the "issues resolved" measure, it doesn't matter when they were created as long as they were resolved in filtered period.

It could be that for your epic (in rows) there were just 6 children (or 5 children + epic itself) created between Jun 2021 and Nov 2021, but 12 (11 children and epic itself) were resolved in the same period. I would guess that some of the children for the epic were created before Jun 2021 but they were resolved in the selected time period.

Martins / eazyBI

[Nick Brown]

Ah ok - yes that makes sense. I have removed that element from it and can now see what you mean :) 

I now have the Issue name twice in my table though:

is there something I'm doing wrong? 

[Martins Vanags -eazyBI-]

@Nick Brown 

I think you need to use just the Epic Link dimension in report rows.

Then create new calculated measure using this formula:

[Issue].[Issue].Getmemberbykey(
[Epic Link].CurrentMember.key
).get('Resolved at')

It will return the epic resolution date when "Epic link" alone is used in the report, because "Issue resolution date" (predefined measure) wont work with "Epic link" dimension.

[Nick Brown]

Perfect - got it all working now, thank you!