Search for: All records

Unlike many STEM fields, data science has emerged with online communities serving as prominent spaces for professional development and learning. This paper explores factors that contribute to gender differences regarding perceptions of satisfaction and difficulty in a learning initiative for data science hosted by the Kaggle community. We investigate multiple factors: prior experience and skills, professional role, and communication within a learning community. Our results, based on a survey of 2,707 aspiring data scientists, suggest that learners who identify as women do not perceive assignments to be more difficult than men, but complete fewer assignments. The increasing difficulty of the learning experience affected all learners, but men were still able to complete the hardest assignments at a higher rate than women despite experiencing similar barriers. Overall, the findings demonstrate how learning initiatives in technically intensive domains contribute to different outcomes between groups. 

Design artifacts in online innovation communities are increasingly becoming a primary source of innovation for organizations. A distinguishing feature of such communities is that they are organized around design artifacts, not around people. The search for novel innovations thus equates to a search for novel designs. This is not a trivial problem since the novelty of a design is a function of its relationship to other designs, and this relationship changes as each design is added. These relations between artifacts affect both consumption and production. Moreover, these relations form a landscape whose structure affects the emergence of novelty. We find evidence for our theorizing using an analysis of over 35,000 Thingiverse design artifacts. This work identifies the differential effects of different forms of novelty, visual and verbal, on subsequent innovation, and identifies the differential effects of different degrees of structure in the landscape on novelty.