I get your point, it is like making a poll about peopleâs favorite food. You add like Steak, Ice Cream, Pizza and then Broccoli. Donât think many people will choose the latter.
The reason I did so was that I wanted to be fair towards the people coming with suggestions. Also, I didnât want to interfere with how data is presented to me. If I group them, or remove some, I am also tempering with the data, which is bad.
I made an excel sheet and listed them individually and then grouped some if there was overlap. Based on that I got 3 distinctive categories. From a qualitative data analysis perspective it makes most sense to run with this. And, simply accept the limitations.
I get that Key Bindings donât have that ââWOWââ factor, so you can kinda tell where it is gonna go 
The bigger limitation I would say is that Key Bindings got added right when I made the poll 
So technically, I could have removed it. But, I felt like it wasnât completely fair to the people who brought it up as some Key Bindings are still not added.
I struggled with this a lot internally. There are limitations to either way. From a logical perspective you just group them all together right. But, it isnât quite ââethicalââ to do so. Since now, you are grouping categories that shouldnât be grouped. Due to the N, the mentions, being too different. It isnât ââriggedââ, it is actually fair, from a statistical perspective. But, I get that it feels that way.
I will say this though, I kinda wanted to yeet Key Rebindings out of it, and group the other 5. But, I didnât feel right about it.
Sorry for the lengthy post. But, I just wanted to give some insight to why I did things. I completely get your point though 
Edit: another issue is that when I group them all together and say these are the most popular. The N difference between the top 3 and the next 3 is almost 100%. And the same was also true for the bottom 33 mentions. Most had 1 suggestion so the average N almost doubled every time. From like 1,3 to 3,3 to 6,7.