And the final speaker in this session at the 5th Data Donation Symposium at the Weizenbaum-Institut in Berlin is Joshua Claassen, who begins by highlighting the ongoing development of data donation approaches and mechanisms. But once data are collected, what do we do with them: how do we transform them into actual findings? There is a need for more standardised workflows in working with such data.
This project drew on data donations from Netflix users; such data packages contain personal viewing histories and other details, but also need to be connected with further information about the content viewed, and such information can be drawn from online sources like Wikipedia and the Internet Movie Database. This also helps to disambiguate between shows and episodes with identical names.
Following such enrichment, it is then possible to aggregate viewing patterns per data donor, exploring preferred genres, viewing durations, viewing frequency, viewing completions, etc. But links with metadata sources are not necessarily always accurate, and this introduces potential errors to the analysis; probability-based linkages could be used to address this, but are inherently unreliable.
