Forum Discussion
Yet another ranking question
I have a large table of data, separated by subject areas (100 rows per subject). The data has been scored by subject according to two scoring methods. Here's a very abbreviated sample:
| Subject | Institution | Score 1 | Score 2 |
| S1 | I1 | 100 | 30 |
| S1 | I2 | 500 | 80 |
| S1 | I3 | 6 | 55 |
| S1 | I4 | 32 | 31 |
| S2 | I1 | 26 | 55 |
| S2 | I2 | 75 | 66 |
| S2 | I3 | 81 | 54 |
| S2 | I4 | 96 | 23 |
| S2 | I5 | 55 | 86 |
| S3 | I2 | 44 | 71 |
| S3 | I3 | 36 | 55 |
| S3 | I4 | 445 | 38 |
| S3 | I5 | 28 | 12 |
I'm trying to create a graph that will allow me to display this data by "Top N" institution. BUT, I want to show "Top N" rankings according to both ranking systems simultaneously *in the same graph*, so that if an institution is ranked highly according to one score but not another, it still shows up in the graph.
Using the above example, if I showed the top 2 institutions according to Score 1, I'd like to simultaneously see the top 2 institutions according to Score 2 in the same graph, even if one institution didn't rank in the top 2 for both Scores.
Is there a visualization that will allow me to do this?
Hi keknight,
Please create below measures:
Sum1 = SUM(Institution[Score 1]) Sum2 = SUM(Institution[Score 2]) Ranking1 = RANKX(ALL(Institution),[Sum1],,DESC,Dense) Ranking2 = RANKX(ALL(Institution),[Sum2],,DESC,Dense) Top2 = IF([Ranking1] <=2 || [Ranking2] <=2,1,0)
Add measure [Top2] into visual level filter, set its value to 1.
Best regards,
Yuliana Gu
3 Replies
Hi keknight,
Please create below measures:
Sum1 = SUM(Institution[Score 1]) Sum2 = SUM(Institution[Score 2]) Ranking1 = RANKX(ALL(Institution),[Sum1],,DESC,Dense) Ranking2 = RANKX(ALL(Institution),[Sum2],,DESC,Dense) Top2 = IF([Ranking1] <=2 || [Ranking2] <=2,1,0)
Add measure [Top2] into visual level filter, set its value to 1.
Best regards,
Yuliana Gu
There could be a visualisation that helps but can you please provide more detail on your example.
Such as what your expected output might look like in data form?
Sure! I'm able to do exactly what I want using Pandas (Python). See above example, where some datapoints have both scores plotted, and some only have one.
