Invisible Data Work : Subjectivity in Decision Making in Data Cleaning

Sammanfattning

Data are being generated at an unprecedented rate, and the magnitude of available data is increasing day by day. This rapidly growing data presents new opportunities for businesses, and the need to analyze it quickly becomes essential. The focus on big data has increased with the rise of artificial intelligence (AI) and machine learning due to their reliance on large datasets to train models and improve predictive algorithms. Despite that data cleaning is critical, it is oftentimes an overlooked work in organizations. Data cleaning is not purely an objective and technical work; it involves interpretive decisions and sensemaking about what counts as “error,” “noise,” “outlier,” or “valid data. This qualitative study explores data cleaning as a form of invisible data work, focusing on practitioners’ subjectivity in decision-making during the data cleaning process. Guided by an interpretivist paradigm, data were collected through semi-structured interviews with data practitioners from different fields. Thematic analysis was used to identify patterns and themes to capture the subjective and interpretive dimensions of data cleaning practices. The study was framed with Weick’s sensemaking theory as a theoretical lens. The findings advanced a new understanding of data cleaning as an interpretive, socially embedded, and sensemaking driven practice and highlight how meaning is constructed at the very foundation of datawork.