argia.eus
INPRIMATU
Umap dies, Twitter's warehouse is over
  • Umap Twitter’s Euskera store began operating in October 2010 with CodeSyntax. In 2023 he died. Elon Muskiz has murdered even though he does not know who we are and who we are. Umap has served as the basis for various services, has for several years rankings and collections of txiolaris, media and traols in Basque; has served to perform sociolinguistic studies, has served to automatically add commented news in Sustatu, has served as the basis for identifying the most shared videos in Basque... All these services have also died.
Sustatu 2023ko ekainaren 30

The Umap bird finder started in October 2010. He has worked for twelve and a half years and has been the group of lice that has gathered 77,132.,076. Just under half, 28,618,588 in Basque, of 27,019 users.

Archive of the interview in Basque

The tweets in Basque that we detected and analyzed year after year have been:

Year Quantity

----------------------

2007 -> 1.375

2008 -> 2.816

2009 -> 20.009

2010> 51.105

2011 -> 181.516

2012> 849.758

2013 -> 2.328.085

2014 -> 2.712.375

2015> 2.809.217

2016 -> 2.791.263

2017> 2.761.630

2018> 2,525,394

2019 -> 2.536.645

2020 -> 3.111.935

2021 -> 2,727,630

2022 -> 2.280.831

2023 -> 927.057

Although Umap starts in 2010, the database also has earlier tweets. How? when a new user is detected, a method to decide whether he was Euskaldun or not, has been to ask for 200 tweets ago and check whether there was Euskera in them. The history of some users of the early years was already there, in those 200 tweets. These collections and classifications have always been programmed.

Each year we have analyzed the use of Euskera with these data, publishing reports. Traol count, count and analysis of the most shared URLs (source analysis) have been performed.

Monitoring of information and news exchange

From the analysis of the links or URLs deployed in Txio, a new service was invented: the automatic informative, which was integrated in Sustatun in August 2012, and was later renamed the Network. This worked like this.

  • Through umap, there were links in the Basque Txios, analyzing links.
  • Acquiring some of its content, which is called a snippet with an image capture, and that was also decided in Basque.
  • When a certain link exceeds a minimum number of tweets and an important algorithm, publish automatically in Sustatun.
  • Among them, some, revised by the editor, bring the surface.

Thus, 7,334,784 links were analyzed, based on 24,901,637 tweets in Basque. Of these, 32,247 news releases to Sustatu, year after year as follows:

Year Quantity

---------------------

2012> 1.135

2013 -> 4.155

2014 -> 3,836

2015 -> 3.962

2016 -> 4.275

2017> 4.119

2018 -> 2.904

2019 -> 1,792

2020 -> 2.704

2022 -> 1.344

2021 -> 1,754

2023 -> 267

Each of these news features added how users were commented to see tweets chains.

A whole era of shared videos

As an extension of the previous service, with Umap we also saw that the videos were increasingly present in the shared content, in the links. So we started collecting on Youtube video (as this platform had a proper API, unlike others) and in January 2017 we launched the service TBX.eus.

Thus, almost 50,000 videos were detected and analyzed, determining that they had content in Euskera, which then exceeded monitoring/sharing parameters that went to the file and were organized according to the ranking of the most inspected. There are 36,727 videos saved in the TBX file between 2017 and 2023. For example, in July last year, what was the most seen in Basque on Youtube? This one.

With the Umap stop in March 2023 he also stopped, what happens is that he has continued to automatically upload the content of some Youtube channels... But without social supplements, without shared data, we will also have to rethink the continuity of this service.

Monitoring

The stop occurred on March 14, 2023, when Twitter closed its APIs open. The last interesting tricks of the day are frozen on the cover of Umap that day.

Since then we have worked at CodeSyntax in a number of technical studies. Did the new payment API conditions deserve effort? We have come to a negative conclusion. Under the new conditions of the Twitter API, we would need a Pro account to be able to continue doing what we have done every day for 12 years, at a rate of $5,000 per month.

If, for academic reasons, for example, we justified a request to go back to work, that would also be pointless. In June, academic APIs have been reduced and Twitter has offered the same to social scientists and data collectors using it: Rate $5,000.

70 million tweets gathered, 33,000 news in Basque commented, 36,000 videos classified by ranking and dates... Is it worth keeping them? Yes, no doubt, either as a thick database for future files, or as a query tool, although Umap and TBX.eus have been closed or frozen, we are committed to storing your content. From now on we will try to organise it well.

Meanwhile, in the case of Sustatu, Twitter has made things even more difficult since the March stop: in recent weeks they have disabled the automated tweets delivery system and the login form for users. We will also have to resolve them.

Twitter, it has been nice all the time and it has not been a pointless job. Don Elon, go to hell.