There are three challenges in using social media content for reporting, as Storyful’s Mark Little has written: finding it, verifying it, and figuring out the best way to publish it.
In breaking news situations, reporters often rely on text searches — names of places, keywords like “crash” or “fire,” and hashtags. They look for users whose bios mention a particular location.
But it’s hit-or-miss. Even when they use the right terms, they have to wade through all the conversation from people who aren’t at the scene.
Geofeedia, a service that comes out of private beta today, aims to solve this problem by enabling location-based searches for social media content. Users can type in a place name, address, even the name of a sports venue, or they can simply outline an area on a map. The service will display the latest geotagged content — from Twitter, Instagram, Picasa, Flickr and YouTube — within that area.
“Most news happens at a location,” said Phil Harris, CEO of Geofeedia. When time is of the essence, “filtering through an unbelievable number of social media posts, it’s daunting.”
Hear about a shooting at a high school in near Cleveland? Draw a circle around the area on a map and start looking at what is being posted, pretty close to real-time. From there, you can filter by keywords and time.
For several months people at about 15 major news outlets around the world, plus others at smaller organizations, have been using the service and offering feedback, Harris said.
Geofeedia (previously Geofeedr) unveils a revamped user interface today, and announces a preliminary pricing structure: $1,450 a month for five users.
The price tag may be an obstacle for local news organizations, but for the time being, the company still is allowing free trials. “We don’t want to discourage people to use it because if they use it, we hope they see value in it,” Harris said.
“Business intelligence” for journalists
It’s easy to see how reporters could use the tool, particularly in breaking news. Besides being a tool to find sources, Harris described it as a form of “business intelligence” that can help editors make decisions about whether and how to cover stories.
For instance, they could use Geofeedia to gauge how big a story is so they can decide whether to send a reporter. If they do send someone, the reporter can go to the place with a high concentration of tweets rather than heading to a general area like a highway exit. And before anyone gets to the scene, people in the newsroom can use Geofeedia to find eyewitnesses and contact them, perhaps while they’re still at the scene of the breaking news.
Harris saw how useful this service was during coverage of a shooting at Chardon High School, near Cleveland, on Feb. 27. He learned of the shooting soon after it was first reported, pulled up a map of the area on Geofeedia and drew a circle around the town of Chardon. He saw a concentration of tweets at Chardon High School and drew a tighter boundary around the school.
“Within 30 seconds of our learning of the shooting, we had identified several sources within the school itself,” he said.
Later, he saw a reporter searching for sources on Twitter, wading through tweets and retweets and asking people if they were at the school. With Geofeedia, “a reporter would have substantiated [someone's] presence in the school immediately and realized that it was a high-priority follow-up.”
How does it fare in real news situations?
I tried it myself on Friday. When I saw a tweet from Reuters’ Matthew Keys about a fire in a high school near Portland, Ore., I logged in to Geofeedia and searched for Portland.
It turned out the fire was at Woodburn High School, south of Portland, so I located that on Google Maps (Geofeedia didn’t recognize the high school name) and drew a circle around the area.
I didn’t have much luck at first. I scrolled through irrelevant content like YouTube videos of homes for sale and Instagram photos of flowers, shoes and, oddly, someone who likes to take photos of handguns.
About 20 minutes later I found a relevant Instagram photo, but it turned out to be a cell phone photo of someone’s TV showing live coverage of the fire. Right after that I found something more promising: another Instagram photo showing a ladder truck at the school.
As with the tweets that Harris identified at the Cleveland high school shooting, neither of those photos had the keywords I would have been looking for, such as “fire,” “Woodburn High School” or #woodburnfire.
Meanwhile, I noticed a reporter for the Statesman Journal in Salem, Ore., reporting the story on Twitter. In the time that I found that one photo, he had retweeted a couple of photos, both of which did have those keywords.
One of several tools for finding witnesses and photos
That short experiment meshed with the experience of Holly Moore, a social media producer for Gannett who’s been testing Geofeedia. She told me she’s used Geofeedia in conjunction with other services, but it hasn’t become her go-to tool.
“I think that it has great potential to be a hip news tool,” she told me by email. “I also think that it’s not quite a service worth paying for yet.”
Moore used Geofeedia during the Chardon High School shooting to find a photo of the place where police captured the gunman.
“I liked it a lot for the tornados in Texas, which was a very visual event,” she said, noting that she used Geofeedia to find two of the 20 photos in a slide show of tornado damage photos. “And it also works well in arenas. I loved looking at what was uploaded during basketball games during March Madness.”
However, Moore said she wished it were easier to get that content out of Geofeedia and onto her website. The new Geofeedia user interface includes buttons to share on the major social media networks and to add content to Storify. The company has also responded to users’ requests to create collections of content they find on Geofeedia.
Dorrine Mendoza, now at CNN, told me that she used Geofeedia when she was at the North County Times near San Diego. She found some material related to an earthquake in Mexico and used it after contacting the sources.
After a series of shootings in Oceanside, Calif., “we had discussed the possibility of using it by saving searches … to get a ‘feel’ for community sentiment.” Though Mendoza didn’t end up using Geofeedia that way, I found the idea intriguing. We know Twitter can be used to eavesdrop on conversations; this would be a way to aim your ear in a particular direction.
Reuters Social Media Editor Anthony De Rosa said the service was useful in covering the May Day protests and in finding images related to a plane that was quarantined at Chicago Midway International Airport. “We probably could have used a Twitter search for this but Geofeedr just made it easy,” he told me by email.
Sometimes the service doesn’t pull up as much content as De Rosa would expect to see, which could indicate how much (or little) social media content is geotagged. I suspect that is on the rise. Also, Geofeedia only pulls up tweets with latitude/longitude coordinates within the search area; it does not look at information like Twitter bios.
That means it may miss some content, but it also helps with verification. “A very high percentage [of the content Geofeedia surfaces] are consistent with the location at which they were geotagged,” Harris said.
Solving the discovery problem
Geofeedia is the second tool I’ve seen in the last few months that tries to help journalists solve the discovery problem. (Verification is another challenge altogether.)
In February, I wrote about a project called SRSR, or Seriously Rapid Source Review, that aims to identify eyewitnesses by looking for certain keywords. Though it looks promising, the project is just a proof-of-concept.
At the time, De Rosa told me, “Nobody seems to have cracked the nut on being able to find that tweet from the person who nobody knew prior to that vital piece of information they posted.”
We’ll see if Geofeedia is the nutcracker, and whether news organizations will pay to find those people.