As more and more data has been generated in the automotive industry in recent years, demand for data analyzing tasks increases among the research and development teams. It thus becomes an inevitable phenomenon that information and knowledge exchange between data scientists and domain experts becomes more frequent. The information flow influences the performance of the team and the outcome of the operation, many companies thus try to implement knowledge management to avoid fiction caused by knowledge gaps and in transparent information among different teams. This thesis will take the collaboration between R&A (domain experts) and GDI&A (data scientist) from Ford RIC as a case to analyze their information flow.
To understand what are the factors that influence the development of a structure to manage the information flow, I look into the literature on knowledge management implementation, and the studies about the characteristics of mediums and tools that influence how information flows between employees. The insights from the literature provide different aspects to inspect information flow in the organization, and become the foundation for designing workshops to research on the current situation between R&A and GDI&A in Ford. An iterative design approach was used to analyze the user and context deeper and more holistic in this thesis. The research started from analyzed the teams from their current situation and culture. And found some information is missing between R&A and GDI&A. They are:
- The context behind the request is not shared, which makes GDI&A could not give more accurate support and feel less engaged
- There is not a repository that documents previous cases and their requests. Therefore, GDI&A needed to cope with repetitive requests, and R&A could not find references for inspiration.
To fulfill both of the requirements, the first priority is to build up better principles of addressing requests in each collaboration project first. Once the workflow of each project and its request is standardized and documented, cases and requests can be viewed by different teams in the organization. Considering the information (request and its context) that has its complexity and the working habit from R&A and GDI&A, the result of the research suggest that
- use asynchronous communication to start the request
- a classification of the request types is needed, it helps R&A formulate their request and GDI&A evaluate the situation to adopt suitable strategies
- synchronous discussion (meeting) is needed to confirm the direction and more detailed information
- an open environment that stored the cases and requests enable the data scientist to reduce their work and domain experts search for inspiration
To make the cases and its request access to those who are not in the team, only high-level information should be in the open environment, external links can be provided for more detailed information.
Master Thesis: TU Delft repository