Which Tweets Should You Clean? Rank by Risk

--
canonical_url: https://digital-footprint-health.shop/blog/which-tweets-to-clean-by-risk
title: Which Tweets Should You Clean? Rank by Risk
tags:

• twitter
• privacy

- security

A friend sent me his X archive scan last week. Seven years, 23,000 tweets, 61 red high-risk flags. His first question was the honest one: do I really have to read all of these?

You don't. I've now run cleanups over 50,000 tweets of my own history plus a few friends' accounts, and the pattern is always the same. You never review a timeline linearly. You rank it. Here's the tier system that fell out of that work, and the order that actually reduces risk fastest.

P0: identity exposure, clean today. One test decides it. If a stranger read this tweet, could they reach the real-life you? Phone numbers count even when they appear once, buried in a reply. Personal emails, especially ones built around your real name. Door numbers, street addresses, "waiting for my package tonight" posts that imply you're home. Scanned IDs. A real company name paired with hire and exit dates, which lets anyone reconstruct your work history. Delete the whole tier in the first pass.

P1: pattern exposure, handle this week. No single post identifies you, but fifty of them map your life. Geotag check-ins, the neighborhood you complain about, the commute you describe daily, the gym and hospital that keep appearing, travel plans posted in advance. These usually carry location or address labels. Batch-delete the ones that reveal a routine and you've removed most of the remaining risk.

P2: person-related, easy to weaponize. Complaints about former employers, because the next HR person will search. Screenshots of region or gender jokes, because they travel without context. Old arguments that got quote-tweeted. Even accurate complaints about vendors can bite later. My test: if this got screenshotted into my new team's chat, could I smooth it over? If not, it goes.

P3: embarrassing but harmless, your call. Angsty poetry, 2 a.m. reflections, eight hundred reposts for an idol. No real risk, and they make the account look human. Keep or delete, doesn't matter.

The mechanical part is quick. Export the archive ZIP from settings, drop it into the checker, filter by P0, open each flagged tweet's original text to confirm the label isn't a false positive, then batch-delete. Move to P1 while you're in the zone. Save P2 for a weekend when you can read slowly. Ten minutes for P0 and P1 combined.

What you're actually handing over when you export is a folder of JSON and media files, every tweet you've ever posted in raw form. The checker parses those files locally and never ships them anywhere. That detail matters more than it sounds, because the whole point of ranking is that you shouldn't have to send a decade of posting history to a cloud service in order to clean it up. Archive in, labels out, all on the same machine.

A few gotchas from running this repeatedly. Deletion is per-account and there's no bulk API endpoint that reliably batches it, so the flow loops over flagged IDs and deletes them in small groups, pausing when the platform rate-limits. Deletes can also take time to propagate, so a tweet you removed may still be visible for a few minutes afterward. Neither is a reason to skip the cleanup, but both are worth knowing before you start so you don't panic mid-sweep.

One more thing about the delete itself: the goal isn't a blank account. It's separating what you want seen from what you don't. Most people land on a comfortable endpoint where the dangerous tiers sit at zero and the embarrassing-but-human stuff stays. The priority table below is the same one the report uses internally, so the labels and the order stay consistent whether you're cleaning a small account or a decade-old archive.

One detail people keep missing: deletion runs on your device with your own credentials. The tool lists what to remove, you click the buttons. That's deliberate, it keeps control in your hands and leaves a short undelete window on some platforms if you hit the wrong thing.

Full priority table and FAQ: https://digital-footprint-health.shop/blog/which-tweets-to-clean-by-risk