Bielefeld Center for Data Science
An Holistic Approach to Data Science
Media and job advertisements for Data Scientists often focus on analytical aspects. The daily routine of a data scientist, advising decision making based on data or developing data products, often focusses on data analysis.
At BiCDaS we decided that, for us, an academic approach to data science should be holistic, meaning we take a step backwards and consider the entire life span of scientific data: We start at a problem or question, move on to data collection and preparation. Then only comes the analysis and questions about deriving value/benefit from the analytic results. Not stopping there, we move on to questions about data storage, retrieval, sharing, re-use and linkage.
This holistic approach to the topic is often modelled in "Data Life Cycle". There are different formulations of this cycle; however, BiCDaS members agree on the formulation depicted in the graphic below:
The data life cycle serves as a guiding model for BiCDaS and structures our discussions and strategic measures. It is important to note that the Data Life Cycle does not identify different, isolated stages. On the contrary, in each stage, the investigator needs to be aware of all the stages because each decision could potentially affect another stage.