Geofeedia: Next Generation Crisis Mapping Technology?

My colleague Jeannine Lemaire from the Core Team of the Standby Volunteer Task Force (SBTF) recently pointed me to Geofeedia, which may very well be the next generation in crisis mapping technology. So I spent over an hour talking with GeoFeedia’s CEO, Phil Harris, to learn more about the platform and discuss potential applications for humanitarian response. The short version: I’m impressed; not just with the technology itself and potential, but also by Phil’s deep intuition and genuine interest in building a platform that enables others to scale positive social impact.

Situational awareness is absolutely key to emergency response, hence the rise of crisis mapping. The challenge? Processing and geo-referencing Big Data from social media sources to produce live maps has largely been a manual (and arduous) task for many in the humanitarian space. In fact, a number of humanitarian colleagues I’ve spoken to recently have complained that the manual labor required to create (and maintain) live maps is precisely why they aren’t able to launch their own crisis maps. I know this is also true of several international media organizations.

There have been several attempts at creating automated live maps. Take Havaria and Global Incidents Map, for example. But neither of these provide the customi-zability necessary for users to apply the platforms in meaningful ways. Enter Geofeedia. Lets take the recent earthquake and 800 aftershocks in Emilia, Italy. Simply type in the place name (or an exact address) and hit enter. Geofeedia automatically parses Twitter, YouTube, Flickr, Picasa and Instagram for the latest updates in that area and populates the map with this content. The algorithm pulls in data that is already geo-tagged and designated as public.

The geo-tagging happens on the smartphone, laptop/desktop when an image or Tweet is generated. The platform then allows you to pivot between the map and to browse through a collage of the automatically harvested content. Note that each entry includes a time stamp. Of course, since the search function is purely geo-based, the result will not be restricted to earthquake-related updates, hence the picture of friends at a picnic.

But lets click on the picture of the collapsed roof directly to the left. This opens up a new page with the following: the original picture and a map displaying where this picture was taken.

In between these, you’ll note the source of the picture, the time it was uploaded and the author. Directly below this you’ll find the option to query the map further by geographic distance. Lets click on the 300 meters option. The result is the updated collage below.

We know see a lot more content relevant to the earthquake than we did after the initial search. Geofeedia only parses for recently published information, which adds temporal relevance to the geographic search. The result of combing these two dimensions is a more filtered result. Incidentally, Geofeedia allows you to save and very easily share these searches and results. Now lets click on the first picture on the top left.

Geofeedia allows you to create collections (top right-hand corner). I’ve called mine “Earthquake Damage” so I can collect all the relevant Tweets, pictures and video footage of the disaster. The platform gives me the option of inviting specific colleagues to view and help curate this new collection by adding other relevant content such as tweets and video footage. Together with Geofeedia’s multi-media approach, these features facilitate the clustering and triangulation of multi-media data in a very easy way.

Now lets pivot from these search results in collage form to the search results in map view. This display can also be saved and shared with others.

One of the clear strengths of Geofeedia is the simplicity of the user-interface. Key features and functions are esthetically designed. For example, if we wish to view the YouTube footage that is closest to the circle’s center, simply click on the icon and the video can be watched in the pop-up on the same page.

Now notice the menu just to the right of the YouTube video. Geofeedia allows you to create geo-fences on the fly. For example, we can click on “Search by Polygon” and draw a “digital fence” of that shape directly onto the map with just a few clicks of the mouse. Say we’re interested in the residential area just north of Via Statale. Simply trace the area, double-click to finish and then press on the magnifying glass icon to search for the latest social media updates and Geofeedia will return all content with relevant geo-tags.

The platform allows us to filter these results further the “Settings” menu as displayed below. On the technical side, the tool’s API supports ATOM/RSS, JSON and GeoRSS formats.

Geofeedia has a lot of potential vis-a-vis humanitarian applications, which is why the Standby Volunteer Task Force (SBTF) is partnering with the group to explore this potential further. A forthcoming blog post on the SBTF blog will outline this partnership in more detail.

In the meantime, below are a few thoughts and suggestions for Phil and team on how they can make Geofeedia even more relevant and compelling for humanitarian applications. A quick qualifier is in order beforehand, however. I often have a tendency to ask for the moon when discovering a new platform I’m excited about. The suggestions that follow are thus not criticism at all but rather the result of my imagination gone wild. So big congrats to Phil and team for having built what is already a very, very neat platform!

Topical search feature that enables users to search by location and a specific theme or topic.

Delete function that allows users to delete content that is not relevant to them either from the Map or Collage interface. In the future, perhaps some “basic” machine learning algorithms could be added to learn what types of content the user does not want displayed or prioritized.

Add function that gives users the option of adding relevant multi-media content, say perhaps from a blog post, a Wikipedia entry, news article or (Geo)RSS feed. I would be particularly interested in seeing a Storyful feed integrated into Geofeedia, for example. The ability to add KML files could also be interesting, e.g., a KML of an earthquake’s epicenter and estimated impact.

Commenting function that enables users to comment on individual data points (Tweets, pictures, etc) and a “discussion forum” feature that enables users to engage in text-based conversation vis-a-vis a specific data point.

Storify feature that gives users the ability to turn their curated content into a storify-like story board with narrative. A Storify plugin perhaps.

Ushahidi feature that enables users to export an item (Tweet, picture, etc) directly to an Ushahidi platform with just one click. This feature should also allow for the automatic publishing of said item on an Ushahidi map.

Alerts function that allows one to turn a geo-fence into an automated alert feature. For example, once I’ve created my geo-fence, having an option that allows me (and others) to subscribe to this geo-fence for future updates could be particularly interesting. These alerts would be sent out as emails (and maybe SMS) with a link to the new picture or Tweet that has been geo-tagged within the geographical area of the geo-fence. Perhaps each geo-fence could tweet updates directly to anyone subscribed to that Geofeedia deployment.

Trends alert feature that gives users the option of subscribing to specific trends of interest. For example, I’d like to be notified if the number of data points in my geo-fence increases by more than 25% within a 24-hour time period. Or more specifically whether the number of pictures has suddenly increased. These meta-level trends can provide important insights vis-a-vis early detection & response.

Analytics function that produces summary statistics and trends analysis for a geo-fence of interest. This is where Geofeedia could better capture temporal dynamics by including charts, graphs and simple time-series analysis to depict how events have been unfolding over the past hour vs 12 hours, 24 hours, etc.

Sentiment analysis feature that enables users to have an at-a-glance understanding of the sentiments and moods being expressed in the harvested social media content.

Augmented Reality feature … just kidding (sort-of).

Naturally, most or all of the above may not be in line with Geofeedia’s vision, purpose or business model. But I very much look forward to collaborating with Phil & team vis-a-vis our SBTF partnership. A big thanks to Jeannine once again for pointing me to Geofeedia, and equally big thanks to my SBTF colleague Timo Luege for his blog post on the platform. I’m thrilled to see more colleagues actively blog about the application of new technologies for disaster response.