I have just discovered this table
It seems different to what you see on the main distrowatch webpage?
Is it perhaps out of date?
Does anyone understand the column headings?
1 Like
Well, I’ll just use a numbered list to explain each column.
- This is the number of the distribution in the list.
- Its name.
- Average rating: all ratings added up and then divided by the number of ratings given.
- The most frequent rating given by users.
- All ratings added up together.
Does this help?
1 Like
Based on user reviews
Lots depends on distro hopping or just used. Bit like trust pilot, booking.com
1 Like
Looks like a list of distributions with a rating average and votes. Interestingly, it is not ordered by rating average but number of votes (or ratings).
1 Like
It seems to be the rank by number of people voting
Correct, I think it is a strange way to order the list.
1 Like
No , I think it is just a count of the number of people doing a rating.
I must admit it is confusing at first sight.
I think it is better info than we see on the distrowatch home page.
1 Like
It is biased by being only the opinions of people interested enough to give an opinion.
The last column (how many people were interested enough to vote) might be unbiased.
Are we ever unbiased and how do you rate something or some one.
If you only use one version…
It’s a good 10 years since I used ubuntu, never used redhat but use the different version of mint so my score if I did it would be biased
No not that . It is biased because it is not a random sample of users. We dont know if the average rating would apply to all users.
Yes, every user has their own varied opinion. That is not bias, that is what we are studying.
1 Like
I wonder if we could get all the individual user ratings … then we could see how variable the ratings are.
1 Like
I believe it is ranked by the people leaving the review. I just left a review for Mint. I found out the review has to be at least 500 characters, so no jumping in and just voting.
1 Like
OK so it is the opinions of committed or interested people only.
That seems to me to be a better measure than the ‘hit counts’ they have on their main site.
1 Like
Does anyone know how to extract data (ie numbers) from the distrowatch sites?
1 Like
Well, it is DistroWatch. I have been using that site for years, and sometimes it is off. The ratings are usually OK, but the way in which they are ranked in popularity do not really represent the reality of Linux usage. But I don’t mind it. It’s fun to search through.
3 Likes
Agree. I would like to see if I can extract something more meaningful from their data.
I have discovered I can copy/paste those tables.
2 Likes
Perhaps a different look at the question.
Why you use distrowatch.com?
As I said before occasionally I look at the top 10, perhaps read the newsletter, but as I have to think about going to the site to read, no way of getting it delivered (unless you know better) lazy and not that effected
I use it to lookup brief info about distros. So when I loom at a distro , that generates one hit. I do not do reviews or ratings. … so my hit means someone was looking for info …
What does your hit mean?
I had a first try at extracting distrowatch data.
Here is a histogram of the HPD( hits per day) data for the last 1 month
Sorry the labelling is generic R stuff. This is a first try.
I used log10(HPD) because without logs it is so skewed you cant see anything.
There seem to be 2 populations
- one between 10^{1.0} and 10^{1.7} (about 50)
- one going up to 10^{3.5} (over 1000)
So maybe there are 2 sorts of hits.
One very annoying thing emerged … Distrowatch allow blanks in their distro names. They obviously dont process their data with Linux utilities…shame on them .
1 Like
Like you, I like to look at the top 10 sometimes. But I also like to look where a distro is on the list when I read about a distro I have not hear of before.
The other day ZDnet listed the 5 “Best” Linux distros alternatives to Windows. Excluding Mint for some reason. Of the 5 listed, I had heard of 2 of them. Zorin and Kubuntu. The other 3 distro I looked up on Distrowatch to see where they were on the list.
2 Likes
When you actually click on a distro and cause a hit, what is your reason for doing that.
I would like to know why people make hits… you can see above I am trying to analyse hits data … i need to understand it.
1 Like