Image: Eetu Pietarinen / Yle
Jarno Koponen
Yle News Lab builds a method that makes the power of social media visible
At the heart of what we do is a team of experts that brings together journalists, data scientists and designers.

Today we live in a world where the digital environments and tech giants directly influence the thinking and behavior of government leaders, companies, communities, and individual citizens.

Our job at Yle News is to explain this complex new world through journalistic means and to help citizens understand new forms of power, emerging in digital environments and affecting beyond them. However, social media platforms and tech giants don’t make this an easy job.

That’s why we at Yle News Lab are building a new data-driven method to help journalists. Our number one goal is to assist and empower journalistic news acquisition and simultaneously increase awareness of the benefits of AI and data design in journalistic work and practices.

At the heart of this endeavour is a team of experts that brings together journalists, data scientists and designers. Concretely, we’re developing an expert tool and a collaborative method for multidisciplinary journalistic teams that can be used to look more deeply at the power dynamics of social media platforms.

ylenewslab data kuvitus1

Most recently, we have collaborated and assisted e.g. in Satu Helin's in-depth analysis of the US presidential elections A peek inside the bubble (which also aroused significant international interest and was thus translated into English) and in Mattias Mattila's Instagram investigation Advertisement or not?. Currently we’re working on a journalistic case in which we use artificial intelligence to analyze large amounts of Finnish-language text to explain what is happening in social media.

In the analysis of thousands of data points obtained through fake social media profiles, human and machine co-operate seamlessly. The classification of tens of thousands of images is done with the help of artificial intelligence, with humans making sure that the data and its interpretations make sense. Statistical methods are part of our basic toolkit and additionally we are using and developing natural language processing methods for the needs of news acquisition.

ylenewslab data kuvitus2

Today, we are in a promising beginning of a long journey. Success will require seamless collaboration and systematic efforts to combine journalism, data science and design.

The biggest lessons to date are that 1) in news acquisition, new methods are always created and selected together - everything starts with the needs of a news story. 2) However, data traces and findings can change the very core of the story or its perspective. The initially planned perspective, data set, or AI-powered analysis method might not work at all and the team needs to be ready for agile iteration. 3) A new multidisciplinary “mini-team” is formed around each journalistic case and story. In such a team, it’s very important to focus on building a common language and professional trust. By combining the strengths of each discipline our journalistic insights get sharper. Put together, the team is more than the sum of its parts.

When all the pieces snap into place, journalists, News Lab experts (data scientists, data producers and designers) and Yle Plus-desk’s great datajournalists turn data and insights into a living story: engaging data-driven and visual news journalism in which humans and machines work seamlessly together.

Author Jarno M. Koponen is the Head of AI and Personalization at Yle News Lab. You can ping him on Twitter here.

Lue tästä suomeksi