Transparent Data Supply for Open Information Production Processes

Open Data and Open APIs have been recognized as valuable approaches for society and business. The validity of data driven decision making can be questioned due to inaccurate data and lack of sufficient provenance knowledge. We explored healthcare data entry situations in administrative patient encounter processes. A wide variety of empirical data was analysed by using frameworks from three disciplines: Human Computer-Interaction, Data and Information Quality, and Software engineering. The analyses revealed ambiguity in timestamps that cannot be recognized from a single perspective. More importantly, they cannot be recognized from the limited perspectives of Open Data or Open APIs that focus on the data layer and data recorded to databases. Unless identified, contextual variations are made visible with additional provenance metadata, they will endanger the validity of data and data driven conclusions. In the future, Open Data and Open APIs should be developed towards Open Information by opening current black-boxes with additional provenance metadata. We also developed general requirements for Transparent Data Supply that would solve several current data quality problems. Information production processes capable of fulfilling these requirements could help secondary users to assess the fitness of information assets for alternative purposes.