Every Facebook user can download their messages, posts, and connections. Meta uses this data to sell ads. Researchers could use it to understand how we actually live online. We help you see what's in your export -- and choose to contribute it to science.
See What Your Data Says What's In Your ExportAny Facebook user can request a full copy of their data from Settings. The export arrives as a ZIP file containing every message, post, and connection from your time on the platform. Here is what is inside.
The full text of every direct message and group chat -- with timestamps, sender names, and thread history going back to when you joined.
Every status update, shared link, and photo upload. Your public broadcast history across the life of your account.
Every comment on friends' posts and in groups, plus every reaction. Often the richest and most honest public writing in an export.
Every friend added and removed, with the date of each change. A complete map of your social network over time.
A list of every company that uploaded your contact information for ad targeting. Most users have never seen this list.
Apps and websites that reported your activity back to Meta -- music streaming, travel bookings, shopping, news reading.
IP addresses, location inferences, device fingerprints, mobile carrier, and session cookies logged over years of use.
Everything you ever searched for on Facebook -- people, groups, topics, places. A window into what you were looking for and when.
In 2021, Facebook whistleblower Frances Haugen revealed that Meta's own research showed its platforms caused harm -- and that the company buried the findings. In response, Meta shut down CrowdTangle and restricted the API access researchers depended on. Here is what consented user data could unlock.
How does language evolve across a lifetime on one platform? What are the linguistic markers of life transitions -- grief, parenthood, recovery? How do people code-switch between public posts and private messages?
Can platform behavior predict mental health shifts before self-reporting? How do sobriety and recovery trajectories manifest in digital communication? What does grief processing look like at scale?
How do political events change private speech vs. public speech? What does civic engagement look like over a decade? How does political self-censorship manifest on social platforms?
How often does location tracking produce false positives? How wide is the consent gap between users and the advertisers who target them? How accurate are ad profiles compared to actual behavior?
What percentage of a platform's social infrastructure depends on a small number of high-contribution users? Are care labor and community organizing roles gendered in digital spaces? What happens to relationships when bridge nodes disengage?
How do people manage front-stage vs. back-stage identity over 20 years? How do attachment patterns manifest in digital communication? How do major identity transitions appear in language before they appear in public?
Do data exports actually contain everything platforms are required to provide? How extensive is the data broker network behind a single user profile? How many apps disclose their data-sharing with Meta in practice?
Does the internet expand Dunbar's number or just inflate the denominator? How do connector types -- bridges, pillars, organizers -- distribute across populations? What predicts network resilience after a user disengages?
How do location data and check-in patterns reveal migration trajectories across economic backgrounds? How do social networks restructure after relocation -- and do lower-income migrants lose more ties? What role do digital communities play in maintaining connection across geographic displacement?
What makes this dataset unique
Most social media research relies on public-only data, self-reported surveys, or ethically questionable scraping. This dataset combines public posts, private messages, ad profiles, off-platform tracking, location data, and search history from the same users -- longitudinally, with informed consent. Nothing like it exists.
Upload your Meta data export and get a private, AI-powered portrait. We'll show you your connector type, your emotional arc, and the invisible ways you supported your community. Then you choose whether to contribute anonymized data to research.
Download from Facebook Settings, then upload the .zip here. Never leaves your browser.
Preview: what your report includes
You connect groups that wouldn't otherwise interact. Remove you, and clusters lose contact with each other.
Your most frequent tone across 16 years -- 34% of all scored messages. Early years show excitement; a grief event reshapes your arc; recent years show grounded warmth.
meaningful connections out of 842 -- above Dunbar's predicted 150
across 9 life categories reported your activity to Meta
Interested in the research dataset? Tell us about your work.
Covering data privacy, surveillance capitalism, or digital wellbeing? We'd love to talk.
We built Ordinary Friend because we believe people deserve to see the value they create -- not have it extracted from them again.
Your data is uploaded over TLS, processed on encrypted servers, and stored in an isolated database. Only you can access your portrait.
Your raw data is never stored. We extract patterns, anonymize everything, and aggregate insights so nothing is ever traced back to you. No names, no identifiers, no exceptions.
We will never sell, license, or share your data with advertisers, data brokers, or any third party. Research contributions require explicit opt-in and are fully anonymized.