Your flow application: what to expect

The flow application is what makes a data donation study fit your research question. It is also where most of the preparation time goes. This page tells you honestly what is involved.

What it does

Download file
The participant requests a copy of their data from a platform and selects the file.
Extract
The flow application reads the file on the participant's device and keeps only what your study needs.
Review
The participant sees the extracted data and can remove anything they don't want to share.
Donate
Only what the participant approves is sent to your study on Next.

No flow application works out of the box

Platforms change the contents and structure of their download files regularly and without notice. A flow application that worked last year, or last month, may silently extract nothing today. This applies to ours, to Eyra's and to everyone else's.

So whichever starting point you choose, plan for three things: someone on your team who can read and adjust Python, recent download files from real accounts to test with, and time to test again shortly before you launch.

Three starting points

Our data donation task

Maintained by D3I Updated October 2026

Comes with standard scripts for the most common platforms. Use the script selector to pick the data types you need, or adapt a script yourself.

Best for: Researchers at Dutch institutions
You need: Basic Python to adapt and test
Support: D3I, Dutch institutions only

Outside the Netherlands you can use it freely, but nobody can offer you support on it.

d3i-infra/data-donation-task

Eyra's Feldspar template

Maintained by Eyra Actively developed

A clean template you adapt to your research question. You write the extraction yourself, or hire Eyra to build or review it.

Best for: Researchers outside the Netherlands, or anyone who wants paid support
You need: Python, or budget for Eyra to build it
Support: Eyra, paid, anywhere

eyra/feldspar

A flow application from an earlier study

Maintained by nobody Probably outdated

Published study code is a snapshot of what worked when that study ran. Useful as inspiration and as a head start for a developer, not as something to upload as is.

Best for: Teams with their own developer
You need: Solid Python, and time to update it
Support: The original authors, if they have time

See the list below

Check what it is built on before you start

Our data donation task and Eyra's Feldspar template are not interchangeable, and each is supported only by its own maintainer. A repository's README will tell you which one it derives from. Who helps you with what shows what that means for you.

Flow applications from earlier studies

The most recently updated open-source flow application we know of per data source. The year shows when it was last updated. None is guaranteed to work with today's download files.