SPIDER: Stata module for spider plots (radar plot)

[Carolina Hincapie]

Dear Statalist community,

Currently I am trying to perform a Spider plot with the package spider, using Stata 18.

I was trying to follow up the tutorial that Asjad Naqvi kindly provide in GitHub:

GitHub - asjadnaqvi/stata-spider: spider: A Stata package for spider plots.

But unfortunately, the by() option is not allowed when I try to perform the code

spider Domain_score , over ( Domain_name ) by ( EQIndex_detoriation)
option by() not allowed
r(198);

* Example generated by -dataex-. For more info, type help dataex
clear
input float subject_id byte Domain_score str1 Domain_moment_BorF byte EQIndex_detoriation str19 Domain_name
6184 2 "b" 1 "Mobility"      
6184 1 "b" 1 "Selfcare"      
6184 2 "b" 1 "DailyActivities"
6184 1 "b" 1 "PainDiscomfort"
6184 1 "b" 1 "FearDepressed"  
6184 3 "f" 1 "Mobility"      
6184 1 "f" 1 "Selfcare"      
6184 4 "f" 1 "DailyActivities"
6184 3 "f" 1 "PainDiscomfort"
6184 2 "f" 1 "FearDepressed"  
5926 1 "b" 1 "Mobility"      
5926 1 "b" 1 "Selfcare"      
5926 1 "b" 1 "DailyActivities"
5926 1 "b" 1 "PainDiscomfort"
5926 1 "b" 1 "FearDepressed"  
5926 1 "f" 1 "Mobility"      
5926 2 "f" 1 "Selfcare"      
5926 2 "f" 1 "DailyActivities"
5926 2 "f" 1 "PainDiscomfort"
5926 1 "f" 1 "FearDepressed"  
5800 1 "b" 1 "Mobility"      
5800 1 "b" 1 "Selfcare"      
5800 2 "b" 1 "DailyActivities"
5800 1 "b" 1 "PainDiscomfort"
5800 2 "b" 1 "FearDepressed"  
5800 2 "f" 1 "Mobility"      
5800 2 "f" 1 "Selfcare"      
5800 3 "f" 1 "DailyActivities"
5800 1 "f" 1 "PainDiscomfort"
5800 1 "f" 1 "FearDepressed"  
5731 1 "b" 1 "Mobility"      
5731 1 "b" 1 "Selfcare"      
5731 1 "b" 1 "DailyActivities"
5731 2 "b" 1 "PainDiscomfort"
5731 2 "b" 1 "FearDepressed"  
5731 1 "f" 1 "Mobility"      
5731 3 "f" 1 "Selfcare"      
5731 4 "f" 1 "DailyActivities"
5731 3 "f" 1 "PainDiscomfort"
5731 4 "f" 1 "FearDepressed"  
5690 5 "b" 1 "Mobility"      
5690 5 "b" 1 "Selfcare"      
5690 1 "b" 1 "DailyActivities"
5690 4 "b" 1 "PainDiscomfort"
5690 2 "b" 1 "FearDepressed"  
5690 5 "f" 1 "Mobility"      
5690 5 "f" 1 "Selfcare"      
5690 3 "f" 1 "DailyActivities"
5690 4 "f" 1 "PainDiscomfort"
5690 1 "f" 1 "FearDepressed"  
5644 1 "b" 1 "Mobility"      
5644 1 "b" 1 "Selfcare"      
5644 1 "b" 1 "DailyActivities"
5644 3 "b" 1 "PainDiscomfort"
5644 2 "b" 1 "FearDepressed"  
5644 3 "f" 1 "Mobility"      
5644 1 "f" 1 "Selfcare"      
5644 3 "f" 1 "DailyActivities"
5644 1 "f" 1 "PainDiscomfort"
5644 2 "f" 1 "FearDepressed"  
5474 1 "b" 1 "Mobility"      
5474 1 "b" 1 "Selfcare"      
5474 1 "b" 1 "DailyActivities"
5474 3 "b" 1 "PainDiscomfort"
5474 1 "b" 1 "FearDepressed"  
5474 3 "f" 1 "Mobility"      
5474 1 "f" 1 "Selfcare"      
5474 3 "f" 1 "DailyActivities"
5474 3 "f" 1 "PainDiscomfort"
5474 2 "f" 1 "FearDepressed"  
5466 2 "b" 1 "Mobility"      
5466 2 "b" 1 "Selfcare"      
5466 1 "b" 1 "DailyActivities"
5466 1 "b" 1 "PainDiscomfort"
5466 1 "b" 1 "FearDepressed"  
5466 1 "f" 1 "Mobility"      
5466 1 "f" 1 "Selfcare"      
5466 2 "f" 1 "DailyActivities"
5466 2 "f" 1 "PainDiscomfort"
5466 1 "f" 1 "FearDepressed"  
5394 1 "b" 1 "Mobility"      
5394 1 "b" 1 "Selfcare"      
5394 1 "b" 1 "DailyActivities"
5394 1 "b" 1 "PainDiscomfort"
5394 1 "b" 1 "FearDepressed"  
5394 5 "f" 1 "Mobility"      
5394 1 "f" 1 "Selfcare"      
5394 5 "f" 1 "DailyActivities"
5394 1 "f" 1 "PainDiscomfort"
5394 1 "f" 1 "FearDepressed"  
5300 1 "b" 1 "Mobility"      
5300 1 "b" 1 "Selfcare"      
5300 1 "b" 1 "DailyActivities"
5300 1 "b" 1 "PainDiscomfort"
5300 1 "b" 1 "FearDepressed"  
5300 2 "f" 1 "Mobility"      
5300 1 "f" 1 "Selfcare"      
5300 2 "f" 1 "DailyActivities"
5300 2 "f" 1 "PainDiscomfort"
5300 1 "f" 1 "FearDepressed"  
end

I also have a second issue, and not sure how to solve it:
The labels in the radar are quite weird (1-2-2-2-2 instead of 0-1-2-3-4-5), and I can't edit them in the edition window









Thanks in advance for your help!

[Nick Cox]

Asjad Naqvi is a member here and might respond to pinging. He last was here on 1 March. My impression is that he hangs out much more on GitHub, Medium or Twitter.

I have never used spider, but I note from its help on SSC this request

Feedback

Please submit bugs, errors, feature requests on GitHub by opening a new issue.

For showing means of several such variables, I would tend to use graph dot or graph hbar.

[Carolina Hincapie]

Thanks a lot Nick Cox !

I posted at the same time the request on GitHub, and indeed he already saw it there, not reply yet.

@asjadnaqvi asjadnaqvi self-assigned this 12 minutes ago
@asjadnaqvi asjadnaqvi added the bug label 12 minutes ago
@asjadnaqvi asjadnaqvi added this to StataViz issue tracker 12 minutes ago

[Nick Cox]

Something like

graph hbar Domain_score , over(Domain_name) by(EQIndex_detoriation)

might help.

If you want further command advice to explore that, then post the means

collapse Domain_score, by(Domain_name EQIndex_detoriation)

dataex

Picky detail: For your paper/report./thesis, use the spelling deterioration.

[Carolina Hincapie]

Dear Nick Cox , thank again!

****NoDet-Baseline****

tab Domain_name if EQIndex_det==0 & Domain_moment_BorF=="b" , sum ( Domain_score )

                   |       Summary of Domain_score
       Domain_name |        Mean   Std. dev.       Freq.
-------------------+------------------------------------
   DailyActivities |   2.9071038   1.4013745         183
     FearDepressed |   1.7650273   .95760741         183
          Mobility |   2.6448087   1.2879675         183
    PainDiscomfort |    2.557377   1.1653126         183
          Selfcare |   2.0218579    1.338181         183
-------------------+------------------------------------
             Total |    2.379235   1.3070276         915


****Det-Baseline****

. tab Domain_name if EQIndex_det==1 & Domain_moment_BorF=="b" , sum ( Domain_score )

                   |       Summary of Domain_score
       Domain_name |        Mean   Std. dev.       Freq.
-------------------+------------------------------------
   DailyActivities |   2.0754717   1.1355562         106
     FearDepressed |   1.4622642   .77058454         106
          Mobility |   2.2924528   1.2186408         106
    PainDiscomfort |   1.9245283    .9729495         106
          Selfcare |   1.6509434   1.1384405         106
-------------------+------------------------------------
             Total |   1.8811321   1.0960587         530



****NoDet-F/U****

. tab Domain_name if EQIndex_det==0 & Domain_moment_BorF=="f" , sum ( Domain_score )

                   |       Summary of Domain_score
       Domain_name |        Mean   Std. dev.       Freq.
-------------------+------------------------------------
   DailyActivities |   2.3114754    1.273668         183
     FearDepressed |   1.4918033   .78354543         183
          Mobility |   2.2622951   1.2740216         183
    PainDiscomfort |   1.9945355   .98615285         183
          Selfcare |   1.7103825   1.1851885         183
-------------------+------------------------------------
             Total |   1.9540984   1.1582036         915



****Det-F/U****

. tab Domain_name if EQIndex_det==1 & Domain_moment_BorF=="f" , sum ( Domain_score )

                   |       Summary of Domain_score
       Domain_name |        Mean   Std. dev.       Freq.
-------------------+------------------------------------
   DailyActivities |   3.1037736   1.2105036         106
     FearDepressed |   1.9622642   .99450422         106
          Mobility |   2.8018868   1.1663522         106
    PainDiscomfort |   2.5943396   1.1692757         106
          Selfcare |   1.9622642   1.3087583         106
-------------------+------------------------------------
             Total |   2.4849057   1.2558967         530

[Nick Cox]

That's not quite what I asked for but with some (a lot of) editing it can be made into a dataset. There is massive scope to display the results in different ways depending on what you most want to compare.

* Example generated by -dataex-. For more info, type help dataex
clear
input str20 Domain_name float EQ str1 borf double(mean SD) int freq
"DailyActivities" 0 "b" 2.9071038 1.4013745 183
"FearDepressed"   0 "b" 1.7650273 .95760741 183
"Mobility"        0 "b" 2.6448087 1.2879675 183
"PainDiscomfort"  0 "b"  2.557377 1.1653126 183
"Selfcare"        0 "b" 2.0218579  1.338181 183
"DailyActivities" 1 "b" 2.0754717 1.1355562 106
"FearDepressed"   1 "b" 1.4622642 .77058454 106
"Mobility"        1 "b" 2.2924528 1.2186408 106
"PainDiscomfort"  1 "b" 1.9245283  .9729495 106
"Selfcare"        1 "b" 1.6509434 1.1384405 106
"DailyActivities" 0 "f" 2.3114754  1.273668 183
"FearDepressed"   0 "f" 1.4918033 .78354543 183
"Mobility"        0 "f" 2.2622951 1.2740216 183
"PainDiscomfort"  0 "f" 1.9945355 .98615285 183
"Selfcare"        0 "f" 1.7103825 1.1851885 183
"DailyActivities" 1 "f" 3.1037736 1.2105036 106
"FearDepressed"   1 "f" 1.9622642 .99450422 106
"Mobility"        1 "f" 2.8018868 1.1663522 106
"PainDiscomfort"  1 "f" 2.5943396 1.1692757 106
"Selfcare"        1 "f" 1.9622642 1.3087583 106
end

label def EQ 0 "EQ = 0" 1 "EQ = 1"
label val EQ EQ 

foreach x in Activities Depressed Discomfort { 
    replace Domain_name = subinstr(Domain_name, "`x'", " `x'", 1)
}

graph dot (asis) mean , asyvars over(borf) over(Domain_name) marker(1, ms(Oh)) marker(2, ms(Th)) ///
by(EQ, col(1) note("") legend(pos(3)) b1title(Mean score)) subtitle(, pos(12) nobox) exclude0 yla(1(0.5)3)  legend(col(1))

[Carolina Hincapie]

Nick Cox really thanks a looot!!!

[Nick Cox]

A further refinement is to move away from alphabetical order. Here I use myaxis from the Stata Journal. https://journals.sagepub.com/doi/pdf...6867X211045582

egen Domain_name2 = group(Domain_name), label 

myaxis axis=Domain_name, sort(mean mean)

graph dot (asis) mean , asyvars over(borf) over(axis) marker(1, ms(Oh)) marker(2, ms(Th)) ///
by(EQ, col(1) note("") legend(pos(3)) b1title(Mean score)) subtitle(, nobox pos(9) orient(vertical)) exclude0 yla(1(0.5)3)  legend(col(1))

[Asjad Naqvi]

Dear Carolina Hincapie (and Nick Cox thanks for the ping!)

I already responded on GitHub but general notes for the community:
spider has bee updated to v1.31 to add additional checks to avoid data issues with three major changes:

• Checks added for by() and over() to have a minimum number of categories (3 and 2 respectively).
• Changed raformat() to just format(), a more standard use.
• format() now shows decimals by default

Please install directly from GitHub for now till the package gets updated on SSC.


Syntax for the graph below:

spider Domain_score, by(Domain_name) over(Domain_moment_BorF) format(%5.2f) smooth(0.5)

[Carolina Hincapie]

Dear Asjad Naqvi thanks a lot!!