Research talk:Cloud Services Annual Survey/2019
Add topicLatest comment: 4 years ago by RoySmith in topic Use of significant digits
Use of significant digits
[edit]It's a pet peeve of mine (drilled into me from my engineering school education), that you shouldn't use more significant digits than the data justify. With 108 respondents, it's meaningless to report 25.29% of anything. That should be 25%, or at most 25.3%. It's even odder that some questions are tabulated to 3 significant digits, and some to 4 digits. RoySmith (talk) 15:21, 10 September 2020 (UTC)
- Makes a lot of sense RoySmith, I've edited the numbers to show only 1 decimal digit ([1]). If you feel strongly that rounding everything to integers would improve even more the readability, I can have a look. For the charts and graphs, it is quite an effort to remake all of them and re-upload them to commons, so I won't be doing this any time soon, but will definitely keep this in mind for next edition in a few months. Thanks for the comment JHernandez (WMF) (talk) 14:57, 21 September 2020 (UTC)
- Thanks, I appreciate your indulging my peeve. I probably would have gone with integer percentages, but what you've got now is fine. RoySmith (talk) 15:31, 21 September 2020 (UTC)
- Makes a lot of sense RoySmith, I've edited the numbers to show only 1 decimal digit ([1]). If you feel strongly that rounding everything to integers would improve even more the readability, I can have a look. For the charts and graphs, it is quite an effort to remake all of them and re-upload them to commons, so I won't be doing this any time soon, but will definitely keep this in mind for next edition in a few months. Thanks for the comment JHernandez (WMF) (talk) 14:57, 21 September 2020 (UTC)