Unlocking consumer insights to boost loyalty and business value
Register Here
The Data Science Day 2025 features guest speakers sharing insightful stories of practical use cases, hands-on workshops and thematic round table discussions. For this fourth edition of the Data Science Day, we have created a programme that leaves even more room for networking opportunities and introduces several hands-on workshops.
Kate Sargent will share an overview of The Financial Times’ “AI Transformation Programme”, which has evolved throughout the year and was launched officially in September to create an overarching framework for AI activities across the FT.
Data- and AI-powered products bring unique challenges, especially when launching and maintaining them in production. These include handling uncertainty from live data, managing the non-deterministic behavior of AI systems, ensuring data quality, and addressing model accuracy. Additionally, rapid advancements in AI, as well as legal and ethical considerations, add layers of complexity. This presentation will explore these challenges and share strategies for managing them effectively, from navigating production risks to aligning with stakeholder expectations.
Prof. Stefan Feuerriegel will examine how AI is reshaping media through cutting-edge methods. He will report about news users’ behaviours using unique findings by evaluating ~22’000 A/B tests at Upworthy and show how to identify drivers of user interest. He will introduce how AI methods can leverage differences in user interests to personalize news content through new AI methods (“causal AI”). This can enable media platforms to tailor article recommendations and other interventions (e.g., discounts), thereby enhancing engagement. He will explore how AI – and specifically LLMs – will change the profession of journalists. Do we need to teach prompt engineering? And how does AI change the quality and perception of news storytelling?
Reusable Data Engineering Products for Scalability, Reduced Code Complexity, Increased Standardization with enhanced Data Quality and Faster Time-to-Market.
The workshops and masterclasses are one of the highlights of Data Science Day. They are a unique opportunity to get to grips with a specialist subject we will explore in greater depth in practical, operational one-hour sessions. Two parallel workshops of your choice are on the programme for this first round of sessions.
Breaking down silos for efficient processes using Integration Platform as a Service (IPaaS) improves agile maturity in data teams and automates business processes, with a focus on enhancing efficiency by prioritizing tasks. Using her experience leading Data Teams, Dr Ana Moya will discuss the practical steps to prioritize cross-departmental tasks for data teams based on several key parameters, including business value, to ensure alignment with organizational goals.
In an era where artificial intelligence (AI) reshapes how newsrooms interact with their audiences, the integration of sophisticated tools like the User Needs Model and AI-driven recommendation systems has become crucial. This lecture explores the strategic implementation of these technologies using a deep dive into the User Needs Model and the innovative application of AI, particularly through the development and utilization of Large Language Models (LLMs). We discuss the integration of expert journalistic knowledge with AI capabilities to create systems that identify user needs, thereby refining content strategy and enhancing reader engagement. Further, we will examine how these systems offer detailed recommendations on how to transform news articles written from one user needs perspective to another, enhancing the versatility and relevance of news content. We also address the ethical considerations and the balance required between AI-driven automation and human editorial oversight to maintain journalistic integrity.
Two parallel workshops of your choice are on the programme for this second round of sessions.
Frankfurter Allgemeine Zeitung is developing an audience management system to deliver personalized content and recommended actions in this ongoing project. Fabian and Tim will guide you through their journey from several different scorings and segmentations that the FAZ data team has worked on over the years to one unified system that aggregates these insights to display the optimal content for each user. They will outline their collaborative process to ensure that efforts to improve subscriptions, registrations and user engagement are aligned, preventing conflicting outcomes and fostering a unified approach.
The Workshop’s theme and host are to be announced.
We plan to end this first half-day at around 22.00, so you’ll have plenty of time to socialize with colleagues and expand your professional network. Meet new people, strengthen existing relationships, and share knowledge and ideas with others.
Understanding subscriber behaviour is essential in a landscape where digital revenue growth is imperative for media organizations. Annika Kiefner will present a machine learning project developed to predict churn and inform pricing strategies targeting low-risk subscribers. She will highlight compelling test results that reveal which pricing models and communication strategies yield the best outcomes. You will gain strategic insights and technical learnings from the project, including the methodologies used and key challenges overcome that finally led to substantial revenue increases at the Rheinische Post Mediengruppe.
Claudia Schulz will discuss the development of iterative AI solutions and the evaluation of human-in-the-loop writing tools at Thomson Reuters (i.e., AI that helps journalists write articles). She will touch upon the subjectivity of what “good” means, considerations Thomson Reuters had when setting up evaluations, and the different iteration steps they went through to find the best solution to deploy.
Participants choose from the programme presented before the start of the programme or offered during the Data Science Day. Each round table, lasting one hour, is moderated voluntarily by a participant. The following topics are indicative and subject to change depending on the interest of Data Science Day participants.
1. Aligning business and analytics operations. Data used to inform decision vs tailored
2. Traditional ML applications: churn, pricing
3. Audio applications of AI: What works best for our audiences: synthetic voices, text-to-speech with AI, newsletter applications, user research
4. User-facing RAGs and chatbots
5. Pricing Strategy: From Insights to Impact: Using Machine Learning to Drive Digital Revenue Growth
6. Data Privacy: EU AI Act, legal side of ML and AI
1. Breaking down silos for efficient processes using Integration Platform as a Service (IPaaS)
2. GenAI and User Needs
3, 4, 5, 6: this is the unconference part of the Day when participants can vote on suggestions put forward during the previous day’s evening social function.
