Group comparison view: update boxplot info in clinical tab #4924
Conversation
✅ Deploy Preview for cbioportalfrontend ready!
To edit notification comments on pull requests, go to your Netlify site configuration. |
TJMKuijpers
changed the title
|
@inodb in the third image: I'm not sure what the standard is --> should boxplot (and scatterplot) be one or two words. It seems that both are used when I do a quick google search. |
|
@TJMKuijpers Good catch! Thanks for trying to come up with a solution to this I think the "Table" name here is unintuitive. How about we call it "Summary Statistics"? In the "Summary Statistics" maybe better to not put the words 'boxplot' nor 'scatterplot'. E.g. imagine sharing a link to this table directly. It's unclear what's meant by scatterplot/boxplot. The boxplot stats are already shown as a tooltip on the boxplot itself. It might be better to only show summary statistics similar to e.g. what the R package does here: https://www.reneshbedre.com/blog/describe-function-in-r.html. So just general data stats (i.e. the scatterplot table) and maybe a few others that the R package reports Does that make sense? |
|
@inodb I updated the table to match the descriptive information Pandas or R returns: Table will be updated when the user clicks on "log scale': |
|
@inodb and @dippindots could you have a look at this PR? |
There was a problem hiding this comment.
@TJMKuijpers Overall it looks good to me codewise, just had a few comments about the type, it seems like we have more than one any type, and we should try to create specific types for them.
There was a problem hiding this comment.
Thanks for updating those types, they look great now. I just have one more comment on this, then we also need @inodb to sign off from the product side
TJMKuijpers
force-pushed
the
scatter-statistics-table
branch
3 times, most recently
from
fce176c to
14f05b0
Compare
There was a problem hiding this comment.
Thanks so much @TJMKuijpers !
This looks great!
Few thoughts:
- MAD - should we spell this one out like
Mean Absolute Deviation (MAD)? - "Lower/higher whisker" - whisker looks odd without seeing the chart. Do we need those? Maybe better to remove? The whisker values show up in the a tooltip of the chart itself on mouseover, so not sure if we need them for the table?
TJMKuijpers
force-pushed
the
scatter-statistics-table
branch
from
0639e54 to
dbcb033
Compare
TJMKuijpers
force-pushed
the
scatter-statistics-table
branch
from
dbcb033 to
81f836f
Compare
Update table with log scale Define types in data table Update screenshot Remove whisker info and add full name MAD Update screenshot
TJMKuijpers
force-pushed
the
scatter-statistics-table
branch
from
81f836f to
a1eeed7
Compare
In the group comparison view, when we go to the clinical tab there is the option to show different clinical attributes per group. Those clinical attributes are either visualized as a stacked bar chart or boxplot.
The figure with the boxplot shows a boxplot with all samples as scatter points.
When we set the visualization type to table, we can see that only the boxplot information is shown:
This is quite confusing for the user since the minimum and maximum values are the lower and upper whisker of the boxplot, not of the data points. We should add this information but also add to the current table that these numbers are describing the boxplot.
This would be the proposed solution (image created with test study es_0)
