“My model shows…”: communicating data science to the management

Finding the balance between model complexity and predictive power is an everyday challenge for a data scientist. Even if the perfect tool is found, communication of the model to the management in an intelligible and convincing way is an area where there is always room for improvement. In the talk, I’m going to share 3 stories of my practice: the usecase of an SVM, a Tree-based method and an ensemble. First, I’m explaining why they were the right tool for the purpose, how I used them in Python, and finally how the results were communicated to the management.

Török Ágoston
Data scientist, Synetiq

Ágoston works as a Data Scientist at Synetiq and is a Researcher at the Brain Imaging Centre, Hungarian Academy of Sciences. His main responsibilities include research and development of algorithms for emotion recognition. Agoston with a background in cognitive neuroscience is specialized on working with using machine learning tools to discover strategies in human behavior and patterns in high density biometric recordings. He is also teaching at the Eötvös Loránd University, and has done research projects at several universities, amongst others at UCL, University of Texas, Austin, and Technion.