Power BI - Find the % matching (Bounty 100) What-if analysis before and after

问题

I have a requirement where I have a table like this,

Role              Skills
Developer           C
Developer           SQL
Developer           C++
Data Analyst        R
Data Analyst        Python
Data Analyst        SQL
Business Analyst    Excel
Business Analyst    SQL

And I need to create something like this in Power BI,

Explaining the first result for a Business Analyst in Power BI Visual Table,

From Filter 1 - I have selected Data Analyst - whose actual skills are R, Python and SQL
From Filter 2 - I have selected a new skill (Upskill) as Excel.

So now, he has 4 skills.

So For Business Analyst - Row 1 in Visual Table %Skills without upskilling - Only SQL from Data Analyst skill matches with Business Analyst skills, so its 50% before upskilling.

But after upskilling with excel it becomes 100% of the Business Analyst skills.

He has 4 skills after upskilling, but 2 of them (R and Python), he won't be using in Business Analyst Role.That is shown in the last column of the table.

The Idea is I am trying to show - How much percentage the skillsets of the selected role is matching with another role.

How much percentage the skillsets of selected role is matching with another role after upskilling.

Happy enough to create new tables based on this and re-shape the data however required.

回答1:

The key here is to have distinct unrelated tables for your slicers.

Let's call your original table Jobs.

Create two new tables:

Role = DISTINCT(Jobs[Role])
Skills = DISTINCT(Jobs[Skills])

Now that we have these tables, we can create slicers with them and read the selected values into our measures.

% Skill Match =
VAR SelectedRole = SELECTEDVALUE ( Role[Role] )
VAR RelatedSkills = CALCULATETABLE ( DISTINCT ( Jobs[Skills] ), Jobs[Role] = SelectedRole )
VAR CurrentSkills = DISTINCT ( Jobs[Skills] )
RETURN
    DIVIDE (
        COUNTROWS ( INTERSECT ( RelatedSkills, CurrentSkills ) ),
        COUNTROWS ( CurrentSkills )
    )

This reads in your selected role in the first variable. When we upskill, we read in the other slicer as well:

% Skill Match Upskilled =
VAR SelectedRole = SELECTEDVALUE ( Role[Role] )
VAR SelectedSkills = VALUES ( Skills[Skills] )
VAR RelatedSkills = CALCULATETABLE ( DISTINCT ( Jobs[Skills] ), Jobs[Role] = SelectedRole )
VAR CurrentSkills = DISTINCT ( Jobs[Skills] )
VAR Upskilled = DISTINCT ( UNION ( RelatedSkills, SelectedSkills ) )
RETURN
    DIVIDE (
        COUNTROWS ( INTERSECT ( Upskilled, CurrentSkills ) ),
        COUNTROWS ( CurrentSkills )
    )

The unused skill measure is very similar.

Unused Skills =
VAR SelectedRole = SELECTEDVALUE ( Role[Role] )
VAR SelectedSkills = VALUES ( Skills[Skills] )
VAR RelatedSkills = CALCULATETABLE ( DISTINCT ( Jobs[Skills] ), Jobs[Role] = SelectedRole )
VAR CurrentSkills = DISTINCT ( Jobs[Skills] )
VAR Upskilled = DISTINCT ( UNION ( RelatedSkills, SelectedSkills ) )
RETURN
    CONCATENATEX ( EXCEPT ( Upskilled, CurrentSkills ), Jobs[Skills], ", " )

The result should look something like this:

You can add some logic to hide the role you've selected in the matrix visual, but I'm keeping things simpler here.

回答2:

Hmm I think the key here is a good model.

The Idea is I am trying to show - How much percentage the skillsets of the selected role is matching with another role.

For the first question you need to define the relationships to evaluate the matches and calculate, based on that, a percentage. I would do the following:

Creating the MaxSkillTable:

JobID    JobName               Skill
 1       Business Analyst     Power-Bi
 1       Business Analyst     SSRS
 1       Business Analyst     Excel
 2       Other jobs           Other skills for other jobs
 ...and so on

Lets say now you select Max Mustermann for Business Analyst on your Person table and get the following result:

     Name         Skill    JobID   
 Max Mustermann   Excel      1     
 Max Mustermann   SSRS       1

Now you need to match the result above mit the MaxSkillTable where the JobID is the same. You will get two matches (Excel and SSRS). This would be your first result. After that you can select the maximal skills count for this job (Excel, SSRS, Power-BI = 3). This would be the second result. When you have this both results you can calculate the percentage.

For Max Mustermann it would be 2 / 3 so ca. 66%.

For your second question,

How much percentage the skillsets of selected role is matching with another role after upskilling.

you can simualte an increasement of the count of Max Mustermanns skills. The result for his query were two skills . So 2 + 1 = 3. Now calculate the percentage based on this increasment again. 3 / 3 = 1 = 100% (here you need to take care to not increase the count over the maximum).

来源：https://stackoverflow.com/questions/55935908/power-bi-find-the-matching-bounty-100-what-if-analysis-before-and-after

标签

powerbi

ssas

dax

data-modeling

powerbi-desktop