New Tool Lets You Donate Your Online Data to Science Safely

Researchers who study how people behave online have a data problem. For years they relied on public posts they could pull straight off social networks. But platforms have steadily shut those doors, closing the access tools that made large studies possible. A team led by Munmun De Choudhury has now published a new system called CANDOR (Collecting and Analyzing Networked Data for Open Research) that takes a different route: instead of scraping what's already public, it asks people to hand over a chosen slice of their own digital history.

Here's how it works. A participant decides which platforms to include, which kinds of data to share, and which time period to cover. That could mean one year of messages rather than everything you've ever posted. CANDOR then sorts through exports from different apps, which all format data differently, and strips out identifying details before anything is stored. Researchers get protected access to the cleaned-up information, and it can be linked to surveys the same person filled out — something public posts alone cannot do.

The catch is that this only works if people volunteer, and asking for private messages or photos is a big ask. The paper is upfront about the tension: researchers sometimes need the actual words or images to study a topic like mental health, but richer data means more risk. CANDOR tries to manage that with participant control, de-identification, and restricted storage, but the trade-off doesn't disappear.

The bigger shift is about who decides. Under this model, you do — you opt in, you set the boundaries, and you know what leaves your phone. If it catches on, studies could examine parts of life that public data never touched, from group chats to health-journaling apps. That's a lot more trust to ask of people, and whether they'll give it is the real test.