Duration of the workshop: 2.5 h
Target groups: data scientists and market researchers interested in browsing data and web behaviour and / or machine learning application to market research.
Is the workshop geared at an exclusively German or an international audience? International
Workshop language: English
Description of the content of the workshop:
Cookies, tracking devices and passive measurement tools provide us with a host of new data regarding people's online behaviour. However big that data is, we often find ourselves in need for going beyond the data in order to predict browsing we did not actually observe. Predictive models of online behaviour are useful in a variety of contexts, such as 1/ when we need to generalize results from a tracked sample to a non-tracked sample (e.g. in order to provide information about online habits for survey respondents or for behavioural sampling), 2/ when we need to feel gaps in online journeys (e.g. because cookies only provide us with partial information), 3/ when we need to explain collective browsing behaviour and not simply to describe it (assuming that a good predictive model also has explanatory value).
The workshop will focus on two predictive tasks, particularly relevant to 1/ and 3/ above: prediction of website visits on the basis of sociodemographic and declarative internet usage data by means of nearest neighbour methods and prediction of news consumption with specific topics by means of generative models.
Goals of the workshop:
The goal is to promote the use of machine learning on passive data as a supplement to declarative data, by helping market researchers figure out what they can ask their data scientists, and by sharing with data scientists some ideas and methods to make the best out of this kind of data.
Necessary prior knowledge of participants:
The workshop will be aimed at diverse communities, from data scientists to market researchers and will not require any specific prior knowledge.
Literature that participants need to read prior to participation: None
Recommended additional literature:
Gleeson, J.P., Cellai, D. & alii “A Simple Generative Model of Collective Online Behavior”, Proceedings of the National Academy of Sciences, 111 (29), 10411-10415, 2014.
Van den Poel, D. & Buckinx, W. “Predicting online-purchasing behaviour”, European Journal of Operational Research, 166 (2), 2003.
Information about the instructor:
Denis Bonnay is in charge of data science at Respondi, where he is dedicated to developing research methodologies for the new kind of data market researchers have access to, in particular behavioural data. He is also a lecturer in Philosophy at Université Paris Ouest Nanterre, doing research on logic, philosophy of science and philosophy of statistics. Denis Bonnay was also a founder and director of data science at House of Common Knowledge. A former student at Ecole Normale Supérieure (Paris), he has a MSc in Logic & Foundations of Computer of Science (Université Paris VII) and a PhD in Philosophy of Science (University Paris I).
Maximum number of participants: No specific maximum
Will participants need to bring their own devices in order to be able to access the Internet? Will they need to bring anything else to the workshop? No