The Data Science Expert Group is an international community of practice for news media professionals working in data science, data analytics and data engineering. Members interact and learn together regularly with a shared repertoire of resources: experiences, stories, tools, and ways of addressing recurring problems.
We share case studies, ideas, resources and connections that help spread good practices.
Members are invited to meetups on average once a month to explore operational challenges around practical business cases. The Group’s online forum helps create a community of practice that lasts beyond regular group meetings, sharing valuable resources and sparking collaboration and discussion between members.
ELIGIBILITY & COMMUNITY GUIDELINES
- Members work in the domain of data science, data analytics, data engineering, data “leadership” or in a related domain. They are news media professionals actively engaged in the evolution of data analytics and actionable metrics to help publishers and newsrooms understand what their audiences value, and what engages them.
- Members of the Data Science Expert Group work for a news organisation member of WAN-IFRA. You are not yet a member, or you’re not sure if you are a member? Fill in the form “Join the Data Science Expert Group” below, and we’ll get back to you immediately.
- Members are willing to share experiences with industry peers to grow common knowledge. They commit to providing at least one contribution per year on the occasion of the Group’s working sessions and roundtable discussions. We are running our meetings under “Chatham House Rule”: participants can use ideas from the sessions, but they are asked not to attribute specific comments to other participants.
to receive your personal invitation, please email us at email@example.com
Personalisation of News with DRIVE. DRIVE stands for Digital Reader Revenue initiative. Designed by the German news agency dpa and the German publishing consultancy Schickler, already ten publishers have joined this innovative way to analyse data, gained insights and accelerated change in newsrooms. They bring with them more than 12 million users and 700 million new data points per month. The technical infrastructure allows comparing the data across the publishers down to the most detailed level, and training self-learning algorithms supporting the newsrooms as if they were one big publisher. Algorithms classify content into topic models that enable the personalisation of the content and more recently the personalisation of the complete customer journey, including individual pricing and personal newsletters. Data and tools are only one side of the coin. The other is the sharing experience. All publishers face the same challenge and hundreds of questions in the day-to-day life of a newsroom. In DRIVE, everything is shared. Not only the data and the analysis but also test results, organisational experiences, successes and failures. REGISTER IN ADVANCE
REPLAY PAST EVENTS
Check the link to the Data Science Expert Group video collection Vimeo showcase
The Globe and Mail’s journey into AI and ML-powered automation, optimization and prediction (April 27, 2022 ). Guest session with Nafid Ahmed, Managing Director, Analytics and Engineering at The Globe and Mail. The Globe and Mail is Canada’s leading newspaper, 175 years old. About a decade ago, The Globe realised that it had to innovate and adapt to the new media landscape. Nafid Ahmed will tell the story about The Globe and Mail’s data-science journey: the obstacles it faced as it started from scratch; their investments, experiments and ideas, and how they resulted in the development of an artificial-intelligence and machine-learning system that is now making money for not only its own publication but also publishers on five continents.
Neue Zürcher Zeitung: Fully Redesigned, Algorithmic-driven, “Next Reads” Section (March 1, 2022). In her talk, Cristina Kadar, Data Science and Machine Learning Product Owner at Neue Zürcher Zeitung, will explain how NZZ completely revamped the next reads section on the article pages. Using an iterative, experiment-driven, and interdisciplinary approach between the product and data teams, NZZ introduced a modular and adaptive layout including several algorithmically-driven article feeds. This resulted in very significant uplifts both in the CTR of the section and in the quality of the user engagement with the content afterwards. Cristina will further go into the algorithmic details of the personalized user feed, a hybrid algorithm developed in-house that is part of the next reads section and has been shown the best engagement downstream from all the data feeds.
News UK: Three tropes for finding insights in your newsroom data (February 9, 2022 ). Modern newsrooms and publishers face an abundance of data, yet for analysts and data scientists, discovering useful insights from this abundance often feels like an uphill battle. Dan Gilbert, SVP Data Science at News Corp, discussed three different methods, which time after time, can help you find those ah-ha moments in the data, or at least uncover less exciting discoveries more quickly. Dan draws from his time at The Times and The Sunday Times to cover practical examples of using these methods, each of which takes their inspiration (coincidentally) from the world of health: # What we can learn from epidemiology when it comes to looking at how many people are visiting our websites and apps. # What we can learn from healthcare when it comes to understanding how people use your websites and apps. # Why context matters, whether you are measuring weight or how many people read one of your stories.
Data Architecture Model at Mediahuis (December 15, 2021 ). Mediahui’s vision and how to identify the need for more quality controls/insight on data and models, approach to data mesh and data quality, the data hub and catalog components, the role of the data steward, data versioning.
The Boston Globe’s analytics approach (October 2021). Ryan McVeigh, Director Consumer Analytics, presents the Globe’s paywall strategy, its offers testing, and churn model during the WAN-IFRA’s Digital Media LatAm conference. Fundamentally, subscriber volume, LTV and retention are not their primary objectives. As strong conversion rates do not necessarily guarantee a valuable income, the Globe makes decisions based on revenue.
Retention is NOT a GOAL. Point and counterpoint debate (October 2021). While retention is extremely important, it cannot be considered as an objective in itself says Louis Deering, Vice President of Digital Subscriptions of MediaNews Group, publisher of the San José Mercury News, Boston Herald, and The Denver Post. Emilie Harkin, Vice President of Consumption Strategy and Growth at The Atlantic will argue that retention is a Goal.
Communicate and Work with data effectively in the newsroom (September 30, 2021). The NZZ’s journey towards developing actionable data insights for the editorial teams, enabling organizational setup, the iterative process of KPIs definition and dashboards development, and how it ties into the daily editorial life.
Dynamic Paywall delivery (June 17, 2021). Anna Kotarba talks about how Gazeta Wyborcza is testing and delivering a dynamic paywall, adjusting content access based on the user’s attributes and the segment they belong to, including the level of engagement, propensity to buy, and churn
Amedia’s Data Strategy (May 2021). Igor Pipkin, Chief Data Scientist, explains how Amedia’s data strategy feeds editorial and business development. this recording was made at the annual WAN-IFRA’s French-speaking forum in May 2021
TAKE OUR “DATA PARTNER POWERHOUSES” QUICK SURVEY
We are inviting you to nominate any company that excels at supplying you with the products and services in your data value chain (analytics, science, engineering). We share your recommendations in our “Data Partner Powerhouses” directory, available to the members of the WAN-IFRA Data Science Expert Group. We hope this directory will bring members inspiration and will guide them in deciding on their future partners.