When Twitter wants to announce a change in how it does things, how should they announce it? With a tweet, naturally: Twitter is removing the ability of its users to precisely geotag their tweets.

Most people don't tag their precise location in Tweets, so we're removing this ability to simplify your Tweeting experience. You'll still be able to tag your precise location in Tweets through our updated camera. It's helpful when sharing on-the-ground moments.

— Twitter Support (@TwitterSupport) June 18, 2019

A tweet may seem like a simple data construct — just 280 characters! — but there’s a sea of metadata sloshing around each bon mot. And since 2009, one of those bits of metadata was the location from which the tweet was posted. The goal? To let you “better focus in on local conversations.” Location was opt-in, meaning it was disabled by default and users had to proactively decide to share that information from their phones. But they only had to decide once, and unless they turned it off, location data would barnacle onto their every tweet from then on.

It wasn’t perfect. It quickly became clear that the dedicated could fake their location. And more recently it came out that, while Twitter would let you tag a tweet with a broad location (“London”), it for years was still attaching much more precise GPS data that, while not publicly visible, could be accessed via the Twitter API. (A nice way to figure out where someone who just made you angry online is right now. 2009 was a simpler, dumber time, privacy-wise.)

Not many people ever used the precise location feature, those who did weren’t always representative of Twitter users more broadly, and given its potential for leakiness, it’s probably a good thing that Twitter is killing it. (I had it turned on by default for a while, but search tells me my last geolocated tweet was two years ago.)

But its loss will be felt to journalists, researchers, and anyone else who used that data to connect the vast expanse of social media to actual real-world locations — which could tell us important things about our neighborhoods, cities, states, and beyond.

During a breaking news event, how can you tell the real eyewitness tweets from some fraudster doing it for the lulz? Looking at a tweet’s location data was one way of helping to verify a real report from the scene.

Want to find other people who were at the location in question, or who frequent it? Searching by location could generate a set of potential sources.

Want to see what people in a particular place are talking about at a given time? Or how one suburb is chatting differently about a news event than another? Geofence the neighborhoods and find out. Want to compare how presidential candidates are being talked about in one state versus another? Want to find out commuting patterns during drive time? Want to see how fans reacted inside a stadium versus at the sports bar a few miles away? Twitter’s location data was sparse enough that you couldn’t count on getting useful data on every question — but often you could, and it was at least a starting point.

The same is true for researchers. Scholars have used Twitter location data to detect where swine flu was likely to spread next in real time. They’ve used it to more reliably estimate crowd sizes and to better understand traffic patterns. They’ve used it to determine who was sharing real photos from Hurricane Sandy and who was just photoshopping from distant dry land. They’ve used it to track the spread of forest fires (and to create one of the all-time best paper titles: “OMG, from here, I can see the flames!”: a use case of mining location based social networks to acquire spatio-temporal data on forest fires). They’ve used it to see what sources of information — governments, media, randos online — people used most during a flood in Pakistan. They’ve used it to show why everyone in Manhattan uses an iPhone but in Newark they’re on Android. (Manhattan’s richer, in case that wasn’t obvious.)

To put it in the context of the dreaded fake news, you can use location data to track where misinformation campaigns are spreading the fastest. (In 2016, disproportionately swing states.) You can even sometimes use it to catch a disinformation campaign in the act. (Woe betide the Russian IRA agent who accidentally tweets out the coordinates of his favorite St. Petersburg deli.)

These were all real and interesting uses based on even the mediocre location data that Twitter users sporadically offered up. And now that’s going to dry up substantially. (There are some caveats: You will still be able to opt in to sharing location data in one context — on a photo taken within the Twitter app. A third-party app can still inject lat/long data programmatically, as some tweets posted from Instagram do. And users will be able to pick a non-precise location tag, like a business, if they wish. But that’s it.)

Don’t get me wrong: It’s a good thing that people have become much more mindful of the information trails they leave online, and it’s a good thing that a tech company has decided to scale back up a source of data that at least could be used for bad purposes.

In a way, we lucked out: Social media’s early days were largely public, usually searchable, and sometimes even useful for analysis. But people are now moving those conversations into private spaces — WhatsApp, Signal, Telegram, group chats. For all the benefits that shift brings, it also makes it harder for people to understand what’s happening in the world around us. Researchers can track (to quite different degrees) misinformation on Twitter, Facebook, and Instagram. But WhatsApp is a black box — and we’ve already seen the havoc activity there can cause in Brazil, India, and elsewhere.

One man’s privacy is another man’s useful data, and that is a very real — if still worthwhile — tradeoff. For journalists, researchers, and others in the telling-people-about-the-world-around-them business, we’ll miss the data when it’s gone.

"I felt a great disturbance in the Force, as if millions of researchers suddenly cried out in terror and were suddenly silenced. I fear something terrible has happened." https://t.co/dlaxFFDGVu

— Nisansa de Silva (@NisansaDdS) June 19, 2019

Dang, so much for linguistic maps based on where people say particular words in their geotagged tweets https://t.co/Dk6omEKkzq

— Gretchen McCulloch (@GretchenAMcC) June 19, 2019

This has just shafted a load of social media managers https://t.co/zD6Lguzl0D

— SecretJournalist (@TheSecretJourn) June 18, 2019

This week in OSINT isn't necessarily a very joyful one.. (via @evanchill) https://t.co/OQaNxRRs22

— Christiaan Triebert (@trbrtc) June 18, 2019