Do you know that Habr has editors? Those who are people. It is thanks to them that the news section is never empty, and you always have the opportunity to joke about heritage .
The editors have dozens of publications per week for each. Sometimes Habr users even suggest that they are not really people, but simply algorithms for finding and adapting materials.
Today we will try to find out how long their working day is, whether they rest at all and whether they have a vacation. Or maybe they are robots? At least some. New detective story on Habré. It will be interesting. Let's start!
Search for victims
It is not difficult to determine which of the Habr users is the editor. They are prolific and write, write, write. Some of them write ordinary posts, others write news, and others write both. Today we will focus on the news. At the time of my initial analysis, the last viewable news page contained publications starting from 03.09.2019/3/04.09.2019. It's December, which means that searching for publications for 04.12.2019 months is not difficult. For an even (actually not) account, I took the period from 4/XNUMX/XNUMX to XNUMX/XNUMX/XNUMX, so that none of the days was only partially included in the data. In addition, a whole week has already passed since December XNUMX, and something tells me that no one will especially read this news. And accordingly, they will not be edited / hidden in drafts.
So, we have 92 days, during which 946 posts were published in the News section. The authors' statistics are as follows:
Rice. 1. Statistics of publications-news
220 publications accounted for , 139 — , 129 — , 122 — and total 86 — . Total - 696 news from 5 authors. None of them is hiding, and in the profile of each it is clearly written that they work for Habré. Another 6 authors wrote more than 10 publications in 92 days, and 19 - more than one. One post in the news was published by 52 accounts.
List of those who published more than 10 news in 92 days
Since we are interested in knowing when editors work and when they rest, those editors who have published the most - the first three - are best suited. After all, I hope that they do not rest, and round-the-clock work will betray anyone.
Let's assume that it's unfair to compare those who work as an editor for several months with those who have been on Habré for years. Or just read all 7.3 thousand posts and 8.8 thousand posts I don't want much. So, , и .
Сбор данных
Since I did not want to manually sort through all the publications a little more than completely, I used automated methods. On the one hand, this deprived the data collection of that warmth and lamp-likeness that is so close to me and always captures consciousness. On the other hand, something tells me that while I re-read or at least leaf through everything written, the number of publications to read can double.
So. The list of publications of each author, available at habr(.)com/en/users/username/posts/ from page 1 to page 20 are recorded. In the next step, each publication is loaded, and the necessary information is written out in one general table of the author's publications.
Acquired information
- publication id;
- date and time;
- name;
- rating (total votes, pluses, minuses, final rating);
- number of additions to bookmarks;
- number of views;
- the number of comments.
This story will only use part of the information, but it would not be very rational to upload posts and not collect everything that is possible.
It is worth noting that from this section onwards, all types of publications are considered, not just news. This is necessary for the completeness of the statistics.
And after a close look at the monitor, you can find a lot ...
The results
1 place
Let's start with the most active Habr editor in the last 3 months. By registering on 26.09.2019/XNUMX/XNUMX, immediately began to write, but never wrote a single comment. The maximum productivity of 6 publications per day was reached 7 times and there were no publications for 15 days. Let's go into more detail now.
Rice. 2. Publication statistics
You may notice that editors have days off. Although, apparently, not every week. The list of weekends can be found under the spoiler. At had a vacation of 8 days at the beginning of November, as well as 3 free Saturdays and 4 Sundays for 80 days. Why vacation and not sick leave, you ask. Sick leave is unlikely to end on Saturday, and on Sunday immediately to work.
Weekend List
05.10.2019/XNUMX/XNUMX (Sat);
06.10.2019/XNUMX/XNUMX (Sun);
12.10.2019/XNUMX/XNUMX (Sat);
13.10.2019/XNUMX/XNUMX (Sun);
20.10.2019/XNUMX/XNUMX (Sun);
02.11.2019/09.11.2019/XNUMX - XNUMX/XNUMX/XNUMX (Sat - Sat);
01.12.2019/XNUMX/XNUMX (Sun);
07.12.2019/XNUMX/XNUMX (Sat).
What about working hours? Posts are published starting from 07:02 UTC (10:02 Moscow time, where the office of TM and Habr is located, if I'm not mistaken) and until 21:59 UTC (00:59). The peak of productivity is from 10:00 to 10:59, and there are quite a few posts before 8:00 and after 19:00.
Number of articles by publication time (UTC)
5 (07:00 - 07:59);
25 (08:00 - 08:59);
27 (09:00 - 09:59);
33 (10:00 - 10:59);
26 (11:00 - 11:59);
20 (12:00 - 12:59);
17 (13:00 - 13:59);
24 (14:00 - 14:59);
21 (15:00 - 15:59);
15 (16:00 - 16:59);
13 (17:00 - 17:59);
10 (18:00 - 18:59);
7 (19:00 - 19:59);
5 (20:00 - 20:59);
2 (21:00 - 21:59).
It is worth clarifying that the working time probably depends on the day of the week, so a few details. For example, on Friday there are no posts after 17:43 - that's why it's Friday. But the latest posts are on Wednesday and Thursday. Details under the spoiler.
Active time (UTC) depending on the day of the week
08:39 - 18:25 (Monday);
07:10 - 19:54 (Tuesday);
07:41 - 21:01 (Wed);
07:02 - 21:59 (Thu);
08:33 - 17:43 (Fri);
07:24 - 17:43 (Sat);
08:36 - 18:27 (Sun)
Since we found out that at least one of the editors definitely has days off (and even holidays?), Let's move on to the most important question. It often intrigues readers of Habr and is periodically discussed in the comments to those posts that they liked the least. Quantity or quality? Do editors have a posting policy?
My answer is yes. Why? Just look at the number of posts per week. With enviable regularity, this indicator fell below 20 only during the rest period, as well as in the first week of work, which was 4 days instead of 7. The average number of publications per week is 23.7, and weekly details are waiting for you under the spoiler.
Number of posts per week
22 (09.12.2019 - 14.12.2019);
22 (02.12.2019 - 08.12.2019);
22 (25.11.2019 - 01.12.2019);
27 (18.11.2019 - 24.11.2019);
23 (11.11.2019 - 17.11.2019);
3 (04.11.2019 - 10.11.2019);
24 (28.10.2019 - 03.11.2019);
25 (21.10.2019 - 27.10.2019);
26 (14.10.2019 - 20.10.2019);
26 (07.10.2019 - 13.10.2019);
20 (30.09.2019 - 06.10.2019);
10 (26.09.2019 - 29.09.2019).
2 place
In second place with 139 posts in 92 days is editor Anya (name from user information). At the beginning of Habr writing on 20.06.2019/255/5, she already has 7 posts on her account. The maximum per day is 66 pieces (reached 178 times), and the most productive day is Wednesday. XNUMX days out of XNUMX were without publications.
Rice. 3. Publication statistics
The number of posts per week ranges from 3 (only once) to 17 (3 such weeks), and the average number of posts is 9.8 per week.
Number of posts per week
12 (09.12.2019 - 14.12.2019);
4 (02.12.2019 - 08.12.2019);
14 (25.11.2019 - 01.12.2019);
14 (18.11.2019 - 24.11.2019);
6 (11.11.2019 - 17.11.2019);
10 (04.11.2019 - 10.11.2019);
15 (28.10.2019 - 03.11.2019);
8 (21.10.2019 - 27.10.2019);
7 (14.10.2019 - 20.10.2019);
13 (07.10.2019 - 13.10.2019);
17 (30.09.2019 - 06.10.2019);
8 (23.09.2019 - 29.09.2019);
7 (16.09.2019 - 22.09.2019);
13 (09.09.2019 - 15.09.2019);
12 (02.09.2019 - 08.09.2019);
4 (26.08.2019 - 01.09.2019);
8 (19.08.2019 - 25.08.2019);
17 (12.08.2019 - 18.08.2019);
17 (05.08.2019 - 11.08.2019);
5 (29.07.2019 - 04.08.2019);
6 (22.07.2019 - 28.07.2019);
3 (15.07.2019 - 21.07.2019);
8 (08.07.2019 - 14.07.2019);
4 (01.07.2019 - 07.07.2019);
13 (24.06.2019 - 30.06.2019);
10 (20.06.2019 - 23.06.2019).
An interesting point with working hours. Publications start at 3:00 UTC and end at 22:33. It seems that someone is slightly reworking, but this is not accurate.
Number of articles by publication time (UTC)
8 (03:00 - 06:59)
7 (07:00 - 07:59);
15 (08:00 - 08:59);
10 (09:00 - 09:59);
24 (10:00 - 10:59);
30 (11:00 - 11:59);
29 (12:00 - 12:59);
30 (13:00 - 13:59);
23 (14:00 - 14:59);
19 (15:00 - 15:59);
20 (16:00 - 16:59);
14 (17:00 - 17:59);
8 (18:00 - 18:59);
9 (19:00 - 19:59);
6 (20:00 - 20:59);
2 (21:00 - 21:59);
1 (22:00 - 22:59).
Which day of the week is the longest? The answer is Friday. In fact, don't forget that I'm ignoring the date and only looking at the day of the week. It is likely that the work schedule just changed a lot. And on September 27.09.2019, 03 at 00:XNUMX, something interesting was clearly happening.
Active time (UTC) depending on the day of the week
07:16 - 19:26 (Monday);
07:29 - 19:37 (Tuesday);
05:11 - 20:17 (Wed);
06:00 - 22:33 (Thu);
03:00 - 20:12 (Fri);
05:20 - 20:31 (Sat);
05:00 - 20:11 (Sun)
Another interesting fact is that this editor also almost never writes comments. 5 comments in 178 days on Habré.
3 place
Final 3rd place for today with 129 posts in 92 days - . In total, he has 359 publications, some of which are still in 2018. When did this user become an editor, or was he from the very beginning? The number of publications has increased dramatically since 01.08.2019/242/1.8. Since then, XNUMX posts have been written, averaging XNUMX per day. Suppose this was the effective date of the powers. So statistics.
Rice. 4. Publication statistics
The most productive day is Thursday and a fairly significant number of publications on weekends. What about working hours? The earliest publication is 02:27 UTC, the latest is 23:25.
A fact that could go unnoticed, but no. 155 publications out of 242 (64.5%) are published in multiples of 5 minutes (:00, :05, :10, etc.). For example, all publications starting from 18:00 are just like that. This happens many times a day. Either someone is very accurate (and has a lot of free time), or articles are prepared as usual, and automatic outputs them from drafts in publication.
In the case of human posts, the amount of time it takes to match this pattern averages 2.5 minutes per article, which is about 387.5 minutes per 155 posts.
For the other two editors, 54 out of 250 posts have this accuracy (21.6%, ) and 54 out of 255 (21.2%, ), which is consistent with the statistics. The decimal system has a 20% perfect chance of encountering a number that ends in 0 or 5.
In this regard, I consider it not interesting enough to study the time of publications. If they are not committed by a person, then this will not provide any information, and if a person, then he has superpowers and it will not be possible to find out anything.
List of most notable XNUMX/XNUMX publications
18:00 - 4 pcs;
17:50 - 4 pcs;
17:30 - 4 pcs;
16:00 - 6 pcs;
15:10 - 4 pcs;
08:40 - 4 pcs;
08:20 - 4 pcs;
08:00 - 4 pcs;
06:40 - 4 pcs;
06:00 - 4 pcs;
05:50 - 4 pcs;
etc.
The time of activity by day also does not betray a real person.
Active time (UTC) depending on the day of the week
03:51 - 23:25 (Monday);
04:00 - 18:30 (Tuesday);
04:18 - 18:20 (Wed);
02:48 - 23:00 (Thu);
04:30 - 17:50 (Fri);
02:27 - 18:50 (Sat);
04:10 - 16:00 (Sun)
Another point that distinguishes him from the other two editors is that he sometimes writes comments. 360 pieces published.
Instead of a conclusion
So, we found out how long Habr's editors work (three of them are the most active news writers in recent times), that they have days off and that some of them are really people and go on vacation.
And we stumbled upon another mystery. Or at least something suspicious. It seems that one of the three listed works in automatic mode, at least sometimes.
Perhaps this is not so. But we have a detective. Anything can happen...
Let's think about this a little more...
And that's all for today. Thank you for your attention!
PS If you find any typos or errors in the text, please let me know. This can be done by selecting a part of the text and clicking "Ctrl / ⌘ + Enter"if you have Ctrl / ⌘, either via . If both options are not available, write about the errors in the comments. Thank you!
P.P.S You may also be interested in my other studies of Habr.
Other publications
Source: habr.com