We got older, we got one and a half times more (we have a third less lovers), and the number of people tolerant of alcohol has grown by a quarter.If you want to see how the Habravites have changed over 5 years, welcome to cat.Background
In the late autumn of 2014, I published an article "Who is Subscribed to Habrahabr?" . It was the result of my exercises with the VK API: I sparsed myself the base of everyone who subscribed to the Habr group in VK, and then briefly analyzed. Gender, country, city, age, etc. etc ... In general, everything that is available from the profile in VK.Habr turned out to be an average subscriber ... However,
I’m better to quote from the comments: A year, two, three have passed ... I periodically reminded myself that it would be nice to write the same article again, comparing the figures of the residents with the samples from the end of 2014.Finally, in February of this (2020) year, I decided that it was time. But then a surprise awaited me.Download Habr subscribers no longer seemed possible.
What to do? To postpone the article for another five years? Well, I do not! The old subscriber base then remained (I will not describe what difficulties it was worth remembering about it, and then getting it).So you can make an equally interesting comparison. Take subscribers from the fall of 2014, download their current data and analyze what has changed in the same people over these 5 years.Base for research
Having obtained the old database and downloaded the user data again, it was necessary to understand, but is there anyone at all with whom to compare?Firstly, a person could delete a profile (I myself deleted it for a long time), and secondly, he could close it, which means that I can’t get the data (it’s also closed for me).
Well and the last: the user could simply cease to be active. On the one hand, so what, the data can be downloaded. But if we are already tracking changes, then it seems reasonable to take those users who still visit VK. Rarely appearing users are unlikely to change the data in the questionnaire (and active is not a fact, to be honest).By active, we mean people who have visited the site at least once in the last 30 days.And this is what happens:
This very top green sleeve is our target audience. They didn’t leave, they didn’t close, they kept activity.Of the 350 thousand that we analyzed 5 years ago, only ~ 154k users have survived to this day (~ 43%!). If you add active but closed pages, you get about 200 thousand or ~ 58%.A similar picture, to be honest, prompts me to think about the fate of social networks in general and VK in particular, but I will perhaps devote a separate article to them. Hopefully sooner than after 5 years.So, we will compare 153,741 users who remain active and open in the vastness of VKontakte. Their data is from 2014, against their own data from 2020 . Go.What has changed then?
Standard note, . , , . / .
« ...». « , , 2014- , ...»
We will not pull and immediately look at how invariable these or other fields in the questionnaires.As you know, not all fields in the profile are required, therefore, we will only compare the changes for those fields that were filled in by the user in 2014 and now.Example.
Enough words, let's get to the numbers.The percentage of field immutability .In parentheses is the number of users who have not changed / changed the field (I recall, we take into account only the completed fields).
You can talk about each field separately, which is what we will do.To begin with, the fields that I will not consider
Sex, university, year of graduationThe most minimally changed parameters. Comments are hardly needed here.Name, religious beliefs, place of workWith your permission, I will skip these fields, since they are all textual, which greatly complicates their analysis. And any typo correction or change of the name “Vanya” -> “Ivan” (namely, this is how the name is usually changed) only clog the sample.Type of employment and level of educationIn order not to delay the article, I will omit these fields. 5 years have passed and it is absolutely clear what happened. Pupils have become students, bachelors are masters, students are graduates and employees of companies. It is unlikely that you can find something interesting here.If you suddenly want to know something about changes in these parameters, write in the comments.And now to the fields, which we consider in detail.Age
To be honest, I thought that basically age change from / to fake dates like 01.01.1905 or something like that.However, no. Let’s take users who have changed their age. Take the delta in days (negative - rejuvenated yourself, positive - aged).The distribution will look like this:
Of course, there are changes for decades, as I expected. But most of the changes are concentrated around zero. Take a closer look, leaving an interval of only ± 10 years.
What are these peaks, I think you guessed it. 365, 730, 1095 days ...People prefer to change their age right away for years, which is nothing.What for? Why change my date of birth to the conditional year 1918, I understand. A kind of demonstration "I do not want to show my year of birth, it's not your business." Reverse metamorphosis is also natural.And move it for a couple of years ... I do not have an exact answer. Just an assumption: if you look closely, they prefer to change to a smaller side, that is, rejuvenating yourself a little.And, yes, they will not accuse me of sexism, but:
The distribution is beveled to the left of both sexes, but there are almost no women on its right side. (If you are embarrassed that men themselves have more values, then this is because their share among the Habra residents is higher).That is, representatives of both sexes like to change the age to a smaller side, but women do it more often.These differences between the sexes are obviously statistically significant (which is confirmed by Mann-Whitney and t-test at p <0.01), but I am extremely surprised by it. It seems strange that someone is trying to hide his real age in VC by changing birth dates.However, it is likely that I am mistaken and this behavior has other reasons than the desire to "rejuvenate."Another possible hypothesis: schoolchildren who registered at age 12 (officially registration is possible from 14 years old), and therefore have thrown themselves a couple of years, and now returning real birth dates back. Of course, I will not check it (at least as part of the article).If you have any ideas, tell me in the comments what else I could miss.Country
Among representatives of the IT sector, migration is quite commonplace.Take the countries to which at least 150 Habr subscribers belong to filter out emissions. And look at:- How many people “left” / “arrived” in the country
- How has the total number of subscribers related to this country changed?
The leader in relative “growth” is the Netherlands and Poland.I can’t say anything about the Netherlands, apparently a good option for moving an IT specialist (yes, my friend moved there a couple of years ago). Why Poland is in the lead, we'll find out a little later.Outflow leader - Ukraine. The result is understandable.Firstly, if I remember correctly, it is now easier for Ukrainian citizens to leave for work in the EU countries.The second direction of possible migration is Russia. The VKontakte social network is blocked in Ukraine and is generally associated with Russia. Considering everything that has been happening between our countries in recent years, it is logical to assume that basically more or less pro-Russian users from Ukraine remained in VK, which could very well have actually moved to Russia.Let's look at where and where they were leaving (I left only the countries where at least 100 subscribers "left" from).
Most of those who left moved to Russia, which is logical, given that VK is a Russian social network. By the way, the smallest percentage of those who left for Russia is shown by Belarusians. But it becomes clear why Poland was in the lead in the previous chart (Ukraine here also contributed).By the way, as for the movements from Ukraine to Russia, one could assume that such a picture was the result of the "transfer" of users from the Crimea. But in the old and in the new dataset there are users from the cities of Crimea who relate both to Russia and Ukraine (I don’t know how VK works here: whether it offers to choose a country or depends on the date of registration, I don’t know). So this is hardly the case.They prefer to "leave" from Russia itself, first of all, to the USA. Next on the list are Ukraine, Germany and Kazakhstan.City
Since we are talking about migration, then cities need to be touched.I will leave only the cities in which the relative changes in abundance modulo were above 10% and in which more or less large numbers of people live (or lived).
Of course, a very tragic situation with Lugansk and Donetsk, I think everything is clear here without unnecessary comments.The fact that the Absolute come in large numbers making it a leading Moscow and St. Petersburg, while in the regions of the negative balance is to be expected, nothing can be changed:
But the leadership of Sochi and Krasnodar relative terms curious. The trend for moving to warmer climes?Surname
If the name, as I said above, is not corrected seriously (basically changing the forms of the same name), then the surnames often change for real. And the reason is obvious: many habravchanka married.
Relationship
Since we are talking about marriage, it is logical to look at the changes in affairs of love. More precisely, in the field of “relations”.Above is the share of users by various statuses of relations and by “was / became” breakdown.Below is the change in the percentage of a particular status in percent. Hereinafter: it became blue, it became orange.
It is logical that we have become older, which means the maximum increase in the status of "married". And this is the only change in the plus. All other statuses went negative. And even lovers.Are there fewer romantics?
But let's not rush to conclusions, let's look at the matrix “It was / became”.
The fewer changes, the greener the diagonal. But this is not so, which is quite logical - for the most part lovers, engaged and having a partner got married (so the romantics fall is fictitious, they just did a level-up).Stability is shown by the status of "married" and free. And half of those with whom everything was complicated solved their problems. And almost with an equally probable outcome: 24% of users became free and in search, they found the same number of partners (engaged, married, married). Where to take the lovers to the end is not clear. Love is unrequited.Political Views
“If at 20 you are not a revolutionary, then you have no heart ...” Further in the text.And what happens in 5 years?Here it is necessary to make a reservation:. ? ? , , ? , .
As we can see, half of the Habrachians still retained moderation or indifference in political matters. Then come the liberals and monarchists.The only significant growth was shown by the libertarian political current. I think this is due to the increased popularity of this movement in recent years. But in absolute terms, it's still crumbs.The matrix this time will be with a clear diagonal.
The most unstable were the Communists. The precepts of Marx and Engels remained true only 3/4. The libertarians have the same, and this despite the growth mentioned above.Which, however, is not so far from conservatives and socialists. The most stable are just the politically moderate and indifferent Habravites.Alcohol
A funny picture, in my opinion :) Over time, it turns out that alcohol is not so bad, and the attitude towards it can be changed to a compromise, if not a positive one.If we consider “sharply negative” as 1 and “positive” ratio as 5, then the average jumped from 2.75 to 2.87 (the differences will be significant, p <0.001).I don’t know what this is connected with. Perhaps the popularization of the wine and beer (craft) culture in recent years has played a role, and alcohol use has ceased to be associated with drinking in saliva. And perhaps, with age, people become less radical in their views and generally cease to perceive something “sharply negative”.Smoking
Finally, the last point.
The same dynamics as with alcohol, but in general, the attitude towards smoking is much more negative than towards alcohol (I noted this fact in the last article, as a whole, a global trend).And again, similar hypotheses: either with age you become more tolerant of everything, or the fashion for vapes / hookahs and other smoking methods, declared as a less harmful substitute for tobacco, has done its job.Conclusion
So that is all. There will be no conclusions, the article is Friday. I just hope that you, a 29-year-old married graduate of Moscow State University, with moderate political views and occasionally sipping wine, was at least a little interesting.