The advent of new technologies has created great opportunities for data-driven discoveries spanning the worlds of science and industry and impacting both equally. In the past few years, data sources and availability have greatly changed. Examples include the abundance of data related to in some way to human behavior, policy implementations, interactions with a large number of computer devices and advertising materials, electronic records, etc. Many datasets are now coming in a range of multimodal forms that are typically collected as streams of information. Naturally, they now present highly unstructured patterns where a single phenomenon of interest is observed through multiple types of measurement devices with each device possibly collecting only partial information of interest. How do these new data types/streams impact statistical and subsequent scientific discovery? How do they change notion of learning and quantification of uncertainty? How can we disentangle and yet utilize all of the complex data that is available in order to enrich statistical algorithms and models? At this two-day conference/workshop environment, we aim at bringing in a group of outstanding researchers that challenge and push the frontiers of statistics as applied to the data science. The conference will focus on the growing impact of data-driven science on the field of statistics and vice versa. Some of the questions that will be addressed are: do algorithms even exist in these new contexts; is their limiting behavior the expected/traditional one; how to quantify uncertainty in scientific discoveries; how to develop efficient algorithms that are robust to the complex nature of the observed data, etc.