Creating distinguishable lines for folks who are colorblind when using graph twoway line

[Jenny Mak]

Hi everyone.

I would love to know how to create different lines by pattern. Below is my dataset.

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(admission_year Type_1 Type_2 Type_3 Type_4)
1  .2255541 .14637904  .1835391      .125
2 .04302477  .1494607  .1654321 .04166667
3  .2346806 .16640987  .2436214 .24621212
4  .1642764 .08474576 .15226337 .24810606
5 .19035202  .1725732  .1876543 .23958333
6 .14211212 .28043142 .06748971 .09943182
end

I used the following code to graph my data.

graph twoway line Type_1 Type_2 Type_3 Type_4 admission_year, clwidth(thick) clcolor(yellow red green blue) title("Admission by Year") subtitle("2012-2018 cohort") legend( order(1 "Type_1" 2 "Type_2" 3 "Type_3" 4  "Type_4")) ylabel(0(.1).5, angle(horizontal)) ytitle("Percentage") xtitle("Year")

Next, I tried to use the msymbol option to distinguish different lines by pattern. Currently, the lines are distinguished based on color but I fear folks who are colorblind may not see the differences. As a result, I ran the following line of code:

graph twoway line Type_1 Type_2 Type_3 Type_4 admission_year, clwidth(thick) clcolor(yellow red green blue) title("Admission by Year") subtitle("2012-2018 cohort") legend( order(1 "Type_1" 2 "Type_2" 3 "Type_3" 4  "Type_4")) ylabel(0(.1).5, angle(horizontal)) ytitle("Percentage") xtitle("Year"), msymbol(D, T, X, S)

However, the code doesn't run. Any suggestions would be greatly appreciated. Thank you.

[Ali Atia]

Code looks fine, just a few unnecessary commas. Try this:

graph twoway line Type_1 Type_2 Type_3 Type_4 admission_year, clwidth(thick) clcolor(yellow red green blue) title("Admission by Year") subtitle("2012-2018 cohort") legend( order(1 "Type_1" 2 "Type_2" 3 "Type_3" 4  "Type_4")) ylabel(0(.1).5, angle(horizontal)) ytitle("Percentage") xtitle("Year") msymbol(D T X S)

Although I don't see the point of the -msymbol- given that this is a line graph. Note the advice in help line:

line, while it allows you to specify the marker_option msymbol(), ignores its setting.

[Jenny Mak]

Hi Ali, Thanks for responding. Your updated code runs but as you mentioned, the msymbol option doesn't do anything. Is there an alternative way to differentiate the four different lines created by your code? For example, is there a way to make one line with diamonds, another line with triangles, another line with X's, and another line with squares? Right now, someone who is colorblind won't be able to distinguish the colors.

[Ali Atia]

You can turn it into a connected scatter:

graph twoway scatter Type_1 Type_2 Type_3 Type_4 admission_year, clwidth(thick) clcolor(yellow red green blue) title("Admission by Year") subtitle("2012-2018 cohort") legend( order(1 "Type_1" 2 "Type_2" 3 "Type_3" 4  "Type_4")) ylabel(0(.1).5, angle(horizontal)) ytitle("Percentage") xtitle("Year") msymbol(D T X S) c(l l l l)

[Jenny Mak]

Thank you Ali. This is exactly what I was looking for. You're amazing. I appreciate you so much.

[Maarten Buis]

Also see this Stata Journal article presenting schemes that allow for people with color vision deficiency: https://journals.sagepub.com/doi/pdf...867X1701700313

[Nick Cox]

See also https://www.statalist.org/forums/for...ailable-on-ssc and use mycolours as a search term in this forum to find examples of its use.