That’s what the verification aid Socialbearing can do

How the analytics tool Socialbearing helps verify Twitter accounts

There are a number of tools that can detect Twitter bots, often by calculating the probability of whether an account is real or not. Such tools are useful for obtaining an initial impression of a user, but they are error-prone: many bots fall through the cracks because their behaviour is too similar to that of real people. Occasionally, genuine accounts are also classified as fake simply because they are very active.

Socialbearing is better than many bot detection tools

That’s why I am a fan of Socialbearing. This free tool is not explicitly intended for bot detection, but it is helpful when analysing Twitter accounts. Instead of relying on an automatically calculated overall rating, Socialbearing lets me keep an eye on all the details myself and draw my own conclusions.

Comprehensive overview of any number of Twitter accounts

After entering a Twitter handle on Socialbearing, you see an overview of the account in the left column of the results’ page. There, you can find information that you can also directly see on Twitter, such as the creation date of the account or the number of tweets. Additionally, further data that are helpful for the verification of the account are analysed.

Score is negligible

Directly under the creation date of the account is the TUQI Score which – according to Socialbearing – indicates the ‘quality of the user’ based on a combination of different criteria. The higher the score, the more likely it is that the account represents a real person.

This score, however, must be treated with caution – similar to many bot detection scores, it can sometimes be off the mark. In my test, very active Twitter users as well as very active bots reached a high score.

Verification aid inclusive

When using Socialbearing, it is best to focus on the information below. For instance, the number after ‘Tweets/day’ shows how many tweets a user publishes on average per day, which allows a user to quickly recognize whether the account tweets an above-average amount, which may be an indication of a bot.

In addition, the Friend/Follower ratio – abbreviated as ‘Frnd/Fllw ratio’ – is valuable information. The number shows the relation between how many other accounts an account follows compared to the number of followers. Many bots follow many more users than they have followers, even though this is no longer necessarily the general rule.

Socialbearing analyses up to 3,000 tweets

Aside from the general overview, Socialbearing analyses the last 200 tweets of the chosen account. With a click on ‘load more,’ 200 more tweets can be loaded if needed.

The tool does not understand the emotional state of German users

Below the indication of the timeframe during which the tweets were published, the reach and the number of retweets and likes, there are three diagrams.

The first is not useful for German users, since it searches tweets for emotion-related key terms exclusively in English to give an indication of the tweets’ common mood. For instance, the German article ‘die’ (the) is misunderstood as the English verb ‘die’ from which the tool draws the conclusion that Germans are often in a bad mood while tweeting.

In contrast, the other two diagrams are suitable for verification purposes. The second shows the relation between the account’s own tweets, its retweets and answers to other users. This way, suspicious accounts that only retweet or that mainly communicate with other accounts can be noticed quickly.

A verification trick that nearly nobody knows

To verify an account, the last diagram, called ‘tweets by source,’ is interesting: it shows which programs and devices were used to publish tweets and thus gives deep insight into a user’s habits and preferences.

This way, it is easy learn that Donald Trump (or his team) mainly tweets from an iPhone, while the candidate for chancellor Martin Schulz uses Twitter from a computer and from an iPhone or sometimes uses Tweetdeck.

The diagram of Frauke Petry’s account is particularly diverse. It is managed from a computer and in equal parts from an iPhone and an Android device. Additionally, she uses Hootsuite and Buffer.

Until now, I have not met many people who know that this information is online; this is an insider tip for the verification of user-generated content since Socialbearing indicates which device or programme was used to post each tweet of the analysed user. For instance, in the following tweet, the specification ‘via iPhone’ appears in yellow.

Verification specialists now only need to ask: ‘From which device did you tweet?’ and they receive another indication of the credibility of their source.

Number of tweets during the course of the year

Socialbearing can do even more, however. On the results page, you also find an overview of the user’s tweet activity over time (see below an example for Martin Schulz), which is useful in finding out whether an account has been active recently or whether it is only active during specific occasions.

The number of Martin Schulz’s tweets from October 2016 to July 2017.

Sahra Wagenknecht’s favourite hashtag is #erdogan

Word and hashtag clouds, which visually present the terms an account most frequently uses, are also useful. However, once again, the word cloud can only be used for English accounts; in German, words such as ‘der’ (the), ‘ich’ (I) or ‘für’ (for) are not filtered, and therefore the clouds often consist of meaningless terms.

The hashtag cloud, however, works well. Below are the hashtags most often used by accounts of German politicians (and also the government spokesman Steffen Seibert, since Angela Merkel does not have her own Twitter account) based on 1,000 tweets:

During a verification process, I would replace the politicians’ accounts with users who have posted relevant, user-generated content. The overview of the most-used hashtags sometimes leads to surprises and helps provide a better understanding of a user.

Sorting a user’s tweets

I can sort a user’s tweets according to different criteria, which leads to even more insights into an account’s Twitter activities.

For instance, the tools allow users to arrange tweets according to the highest or the lowest number of retweets, as well as according the number of likes and their reach. I can also see which tweets have been interacted with most often.

It is, however, a pity that all of an account’s retweets are included. This means that the tweets of an average user only rarely appear at the top of the list because they often do not have the same reach as retweets (for example, from media or well-known personalities). This results in less useful results from an analysis function because, in most cases, I want to learn more about an account’s original tweets.

Are my followers bots?

Another function that should not be neglected is the follower analysis. Here, I can find out more about a specific account’s followers, which is particularly helpful for in-depth analysis of bot networks.

Follower analysis can be found by selecting the button “followers” on the start page and entering a Twitter handle into the search box.

Socialbearing loads 100 followers at a time of the analysed account and shows a general overview (as described above) for each. The followers can be further sorted, for example, according to the number of tweets, the number of followers or the age of the account. By applying these criteria, followers who are likely to be bots can usually be found quickly.


Try it out! Socialbearing is a comprehensive analytics tool that can be used for hours without becoming boring. It helps verify user-generated content and is also free, which is not often the case for a tool with such a wide range of functions.

Leave a Reply

Required fields are marked*