This site is a prototype for a service that researchers could host to collect personal data donations, using OAuth and service APIs to get user consent to pull personal data directly from the services where the data is hosted. For the kinds of personal data available from APIs, this presents an opportunity to streamline the experience of donating data. A participant may be recruited to join a research study because they use a specific service, but they may also be recruited by a clinic, a professor, or any other kind of outreach based on the study's profile for participants. Once recruited and visiting the study's data donation Web page, only a few clicks are needed for the user to log in privately to the data platform and consent for the data to be transferred. The process could take less time than the average online research questionnaire and share only the information that the user consented to.
As a proof-of-concept, the implementation currently has some limitations that could be overcome with more investment.
- Data is not anonymized on receipt (though it could be depending on the research purpose)
- There's no data deletion request implemented yet (though that could be done by email instead)
- Data is pulled one-time, though access tokens may allow repeat access
So far, this prototype has only implemented data requests for services that have OAuth/HTTP APIs and low barriers to 3rd party API access. If the site was hosted by an actual research organization able to apply to services for 3rd party API access, many kinds of data could be available including:
- Messages, social media posts and comments, block lists and follow lists, from social media and forum sites like Tumblr, Instagram, Tiktok, Reddit and more
- Personal Internet usage history such as search history, AI conversations media views, music/podcast listens, ad impressions
- Heart rate or blood-glucose or other health monitor data
- Exercise or food tracking apps like Strava
We're starting to talk about the future of this kind of approach. There's definite benefit to multiple researchers banding together - they can share the expense of securing the Web site and database, and transforming the kind of data they need into what the research studies require. Also it might just be a matter of an extra checkbox to ask for user consent to allow multiple studies access to the data they're donating. A research institution specializing in a particular kind of personal data (e.g. online safety data, or health tracking data) has an opportunity to really benefit multiple research efforts with the same data donation site and pool of participants.
Contact the Data Transfer Initiative and let us know your thoughts or interest!
To run locally, first install dependencies with pip:
python -m pip install -r requirements.txtThen, initialize the local database:
python manage.py migrateFinally, start the local development server:
python manage.py runserverTo access the admin site, create a superuser account:
python manage.py createsuperuserThen visit /admin and log in with your credentials.