On June 19, Agustín Bignú, co-founder of Human Trends, was interviewed on the CNN Economy program directed by Santiago Bulat and Julieta Tarrés.

CNN Radio Argentina - Wikipedia, la enciclopedia libre

In this interview he talked about the beginnings of Human Trends, the services we offer, the reports made about the protests on June 20th occurred in Argentina and how certain practices are carried out at Human Trends.

Interview

Reporter:

Tell us a little bit about who you are and what you do.

Agustín:

“Human Trends is a technology-based company that performs two services: the first is research in social media and open sources such as Twitter, Facebook, Instagram or newspapers, the second is to apply artificial intelligence algorithms to personal data that can be provided by a company, for example”.

“Within social media research, data is collected in social networks and open source media in order to carry out a study. For example, investigating a demonstration or social movement, detecting disinformation and discrediting campaigns (fake news, trolls in social networks), etc. We are able to detect where that attack is coming from”.

“We also do marketing campaigns, market evolution, viral marketing of influencers. Related to this we have a product that is being very acclaimed in Spain that is focused on influencers and how to get their image on social networks. That is to say, its influence at a general level in the social network. This makes them know their followers better and can focus their publications towards the point that they have the most influence and, thus, be able to increase their number of followers”.

Reporter:

How are you able to measure user influence on a social network? Let’s do an example with Santiago, for example, how would you measure it?

“What we would do is run our algorithms on Santiago’s Twitter followers and build a network of connections, i.e. people in whom Santiago has influence. This influence can be direct, if they are specific followers of Santiago, or indirect if they are connected through someone else. In the specific case of Santiago, it directly connects with 70k people on Twitter, but indirectly it can easily connect with more than 1 million people. We are able to tell you which communities are in these followers, which of them are the closest to your posts and also which contacts are the most important to keep close. With all this information that is open in social networks we can make a personalized report so that each influencer knows how to focus their business”.

Reporter:

You are in Spain. Do you make studies adn reports for Spain only? Do you work for other parts of the world? How long ago did you start?

Agustín:

“We started recently, a few months ago. But quickly, thanks to the fact that artificial intelligence and data are a trending topic, we started to touch on different points. For example, social media research is of great interest to state security forces. We are talking with the police in Spain, in some projects, we are also talking with ethics bodies to collaborate in research projects. At the same time, we are talking to different newspapers about analytics, the impact of readers. We are also talking to companies, especially about advertising and marketing on the subject of market evolution, viral marketing, etc.”

Reporter:

“These reports of June 20th protests, did you do it because someone asked them, did you do it because you wanted to? Tell us a little bit about it.”

Agustín:

“We addressed this report because everything that was happening in Argentina came to our attention. So we decided to investigate it.”

Reporter:

“Tell us what topics you studied within this report. And above all the results, to understand a little more what can be found from the networks in society.”

Agustín:

“This study has three different parts: the first is to locate the analysis, what is the social phenomena, why it happens, when it will happen and where, the situation before it happens in the streets since this study consists of two reports, a pre-report analyzing the background and conception of the movement in which predictions are made about what will happen and a post-report that is made to verify all the predictions and get some feedback about them and, above all, about what happened. The second part consists of data analysis. First we extract the data from open sources (social networks, newspapers…), then, with our programs, we are able to map the extracted connections and assemble an aesthetic and visual network that allows us to represent the movement and evolution of it over time. The pre-report covers a week, approximately. Furthermore, with our intelligent algorithms we can detect who the drivers of the movement are, the narrative that the movement has, the discourse that these drivers use, all from a point of view that respects the privacy of each user. Knowing all these data we are able to predict the number of people who will participate in the demonstrations. The post-report, mentioned above, would be the third part of the study.”

Reporter:

“How accurate can that estimate be? Since they may be in favor but perhaps not physically going.”

Agustín:

“We have come to very innovative conclusions which are that the majority of people who are most reactive (reacting to comments, constantly publishing on the subject…) in terms of a social issue on social networks are the most active participants in the movement. When we say “active participation” we refer to caravans in cars, pots and pans, protests in public places, it is an estimate not in one place but globally. It would be an external view of the protest.”

Reporter:

“How do you detect trolls?”

Agustín:

“Imagine a company that has smear campaigns against it, or has people trying to trash their product. What we would do is get a network of that company’s followers and identify those with the most activity within it. We would, in turn, apply filters, that is, remove the noise, followers who are not important until we have a sector where these fake news or trolls come from”.

Reporter:

“Can you tell us about any cases that have surprised you about Argentina?”

Agustín:

“A report was made on protests that took place at the end of May, specifically on May 25. This protest began with small traders claiming the right to reopen their businesses. What began as a protest demanding these rights turn into a political message and, from May 25, it stopped being a protest of businessmen and started being a political movement. I am talking about specific political parties that were involved in the movement’s network. Then, what is surprising is how a protest that apriori claimed for commercial needs due to the quarantine derived in a political protest influenced by some parties of the Argentine politics”.

Reporter:

“If you had databases that didn’t change live, e.g. static data. For example, databases to detect crime trends, could that also be done? Could the real-time data be combined with the static data?”

Agustín:

“Exactly, we would be able to do cyber-patrol, for example. Detect the hot spots of a city and even go as far as to predict criminality in areas of cities to prevent crime. That is, what times are the most dangerous, when is the next crime expected, where, etc.”

Agustín Bignu

Data Scientist – Physicist

Leave a Reply

Your email address will not be published. Required fields are marked *