This was originally posted in December 2015. Much has changed since then, but the core concepts remain the same.
In 2008 while conducting social media analysis for a dignitary visit, we used a rudimentary scraping tool, designed by a couple of bored college kids. It was slow. It was choppy. It was all we had.
Despite its weaknesses, we discovered one Tweet that forever changed the opinion of how social media would impact law enforcement; “The Sheriff is coming. Does anyone have a rifle I can borrow?” Within 38 minutes of the Tweet hitting the network, we located our suspect only yards away from where the Sheriff was speaking. We didn’t have her geolocation. All we had was her comment, her profile information, and her photo. We were forced to fuse old school detective work with the new, unexplored universe of social media. This one incident was a clarion call announcing the war on crime was going to be fought in a place we’d never seen, the cybersphere.
Today, social media aggregation and analysis is a staple in every major city. Officers are dismantling sex trafficking rings. School Resource Officers are preventing teenage suicides. Thousands of fans are able to enjoy their favorite sports all because a small cadre of savvy officers are quickly identifying and quietly removing the criminal element from their midst. The world of crime fighting has truly changed, and it is set to change again.
A pillar of social media monitoring for many years has been geolocation. Entire companies were built on the ability to extract detailed location information from social media users. As the habits of social media users evolve, we are seeing a general trend away from location based data. Security minded users are switching off the “GPS” on their phones and manually overriding the inherent location reports in many apps. Whereas 18 months ago, location rich data filled the feeds of aggregators, it now supplies less than half of the useful data. Exact location information has given way to “check-ins”, or “soft geolocation.”
In an ironic turnabout, social media analysis has reverted back to those younger days in 2008. Savvy analysts now must work to identify key words and hashtags associated with their investigations, and then use them to find the data they need. This is a multi-step process, but once accomplished will reveal troves of data. The good thing is, it is not hard to do, and even a moderate user of an aggregation tool should be able to gather enough information to derive meaning from the data.
In short, you must look at key words and hashtags as parts of a conversation. Imagine yourself at a large party in a hotel ballroom. As you make your way through the crowd, you listen for clues to help you find interesting conversations. Like cars? Then you will listen for words like horsepower, peak torque, and so on. Into soccer? You will listen for words like Messi, Premier League, and Sounders. Once you hear familiar words associated with your interest, you will gravitate to the conversation, and attempt to enter it by using the very words which pulled you in.
Social media is exactly the same. The tools currently on the market allow you to walk through the ballroom, stadium, or school hallway, listening for interesting conversations. Recent ones include #selfharmmm (teenage suicide), #blacktwitter (African American youth), and #qaribanqariba (terrorism). Gathering the words you need builds a lexicon you can rely on to pull content rich data from any event, anywhere in the world.
Geolocation will always be around in one form or another, but for the time being it is on the decline. As social media analysts and investigators we need to build a tool box that will stand the test of time, and be deployable regardless of the trends. Matching our resolve to the ever evolving world of social media will ensure we remain on the forefront of saving lives and protecting the innocent.