14:10-15:10: Democratizing Data Science by Bill Howe
Why does data science remain so difficult? Despite years of focused research in systems and methods, data science remains a high-touch, attention-intensive exercise. There has been a “Cambrian explosion” of big data systems proposed and evaluated in the last eight years, but relatively little understanding of how these systems or the ideas they implement compare and complement one another.
15:25-16:10: Data Science and some (classical) challenges by Søren Højsgaard
Once upon a time, statisticians (data scientists of the old age) often distinguished between data from planned experiments and data from observational studies. They also focused on writing down the questions to be asked (sometimes phrased as hypotheses) in carefully prepared protocols before data was collected. The purpose was, amongst other things, to protect against jumping to unwarranted conclusions based on spurious data. Modern data science together with powerful computers offers new possibilities on the type and scale of problems that can be handled. Perhaps some of the good old deeds should be revitalized in this context?
15:25-16:10: Extracting Value from Big Data – The Case of Vehicular Traffic Data by Christian S. Jensen
Almost all areas of everyday life are accompanied, guided, and influenced by computing and communication devices that are embedded in large networks, most notably the Internet. These devices produce increasing amounts of data. There is consensus that society and businesses can benefit substantially from the ability to base their functioning on these large amounts of data, often called big data. Notions such as “data-driven business” and “data-driven society” have been advanced, suggesting that entities that are able to base their decisions and operation on data are capable of being more competitive than those that are not. Big data is slated to have a profound effect on society. We are in the middle of a revolution of the way we live, work, and interact.
16:10-16:30: Plenary discussion
InfinIT - Innovationsnetværk for it