On March 8th 2021, Connected Data Group took part in the Mediahuis Insight Summit ‘Unlocking the potential of data’, a corporate initiative of one of the leading media groups in Belgium and the Netherlands, to inspire and train their employees.
Mediahuis is the international media company behind a wide range of cross-media brands in Belgium, the Netherlands, Luxembourg and Ireland with 1.4 million newspapers, over 2 million digital news consumers and with more than 3.200 employees that work daily for the (newspaper) titles, radio and television channels and online platforms Dutch media brands include: De Telegraaf, NRC Handelsblad, NRC-Next, Metro, Privé, Autovisie, Groupdeal, Gaspedaal.nl, Relatieplanet.nl, Jaap.nl, Speurders.nl, Dumpert, Culy, Famme and many others. For advertisers, Mediahuis is a strategic partner that provides input for commercial solutions.
‘A meaningful difference thanks to Data Science’
A group of 40 participants saw Erik Fransen (co-founder Connected Data Group and Senior Business Consultant) present the basic principles and business value of Data Science. Data Science can make a difference to those who have, like Mediahuis, a innovative view of their business. With Data Science one can make optimal use of data and finding previously unknown or unexpected insights to leverage the enormous business potential in the media market.
At first Erik stated, “Data Science is not only about statistics, algorithms and complex mathematics. Although it’s a core part of it, Data Science is a complete set of methodology and techniques, focusing on data management, data engineering, machine learning and data mining. And then applying that successfully in the daily practice of running a media corporation.
To apply Data Science and make it successful you don’t have to design the most advanced Machine Learning Algorithms yourself. There is an enormous library of algorithms available to do the finest analytics. Rather, to make the outcomes highly effective the translation from the specific business question to a clear data mining question is key, Erik explained. It all starts with asking the right questions before starting your data science journey.
‘Build your own algorithms’
A deeper dive into different algorithms for Data Science and a Hands-on-Lab with the Data Mining tool ‘Rapidminer’ took place in a separate workshop ‘Build your own algorithms’.
In this workshop with interactive sessions, in four Break Out Rooms, a limited group of Mediahuis-employees learned how to set up a predictive model in the RapidMiner tool based on the recommended CRISP-DM (CRoss Industry Standard Process for Data Mining) process with three provided files, a case description and a business question.
Looking for previously unseen (cross) connections in data, they intuitive explored Data Science without having to have Phyton and R skills.
At the end of the interactive sessions, the following three conclusions were shared:
- Data Mining is the confluence of multiple disciplines, amongst others Information Analysis, Data Mining, Data Management and Visualization;
- Modern data mining tools are easy to use but it’s just a small part of the job. The hard work in Data Science is actually not running the algorithms, the hard work is data preparation and interpretation of the results;
- Data Science is also about how to limit wrong conclusions. You can make a data mining model that fits the data, but you still have to find the best fitting models that really help your answering your business questions.
Want to know more?
Then contact us and talk directly to Erik. He is happy to answer your questions.