Guardian election road trip review: Geo-tagging


View Larger Map

With the inauguration of Barack Obama as president of the United States now well behind us, I thought I’d take (a long overdue) look back at the road trip that I took during the US elections for The Guardian and talk about some of the things we tried in terms of innovations in coverage and what I learned from it.
This is the third trip that I’ve taken for the US elections. In 2000, I took a trip with BBC Washington correspondent Tom Carver. Webcasts were the thing of the day, and we took a portable satellite dish and a DV video camera to webcast live or as-live (recorded but treated as live). We answered a range of questions covering topics suggested by our global audience. In 2004, I took another trip with BBC News Online colleague Richard Greene. The trip was my introduction to blogging, and it set the path for my career for the last five years.

The common thread through all of these trips has been an attempt to engage the audience in new ways and field test new digital journalism techniques. Over a series of posts I’ll talk about some of the things that we did for US Election trip 2008.

Geotagging

As I mentioned last summer, one of the things that I wanted to try was geo-tagging. I was inspired by the GPS and geo-tagging function in my Nokia N82 to add this to our coverage. The camera in the N82 is stellar. With a 5 megapixel sensor and a brilliant Xenon flash, it is one of the best features in the phone. (I’d be interested in seeing what the new N85 has to offer, apart from the OLED screen. ZDNet has a review.) I’m going to focus on geo-tagging in this post and talk more about mobile newsgathering with the N82 and other smartphones in another post.
As good as the camera is on the N82, I knew that there would be times when I needed Suw’s Nikon D70, a proper D-SLR with interchangeable lenses. But how to add the geo-data? Dan Chung, award-winning photographer and digital innovator at the Guardian, and I had played around with a geo-tagging device from Sony, the GPS-CS1.

A geo-tagger at its most basic has a GPS radio and some memory. It records your location either every so often or after you move a certain distance. It’s not physically connected to the D-SLR in any way, but it does require you to sync the clock from the geo-tagger with the clock in your D-SLR. To add the geo-data to your photos, all you have to do is import the photos to your computer and import the GPS logs from your geo-tagger. Software then compares the time that the photo was taken with your GPS logs and merges the geo-data into the EXIF files of the photos. Newer high-end cameras such as the D200 have GPS add-on units (the GP-1), and point-and-shoot cameras like the P6000 have integrated GPS.

Dan had me test the Song geo-tagger a couple of years ago, and I wasn’t that impressed. It didn’t acquire the satellites very quickly, and Sony didn’t officially support non-Sony cameras. But although the accuracy wasn’t brilliant, the idea is sound.
I looked around and settled on GiSTEQ CD110BT. It has a sensitive MTK chipset with 51-channel tracking, and I found the accuracy to be frighteningly good. The GPS track plotted on Google Earth actually shows when I changed lanes in my rental car. The Sony could take minutes to acquire the satellite, but from a cold start, the GiSTEQ usually got a lock in less than a minute. A synthesised voice says “Satellites fixed” when it’s got a lock. To conserve power, it will shut itself off but wake when moved or vibrated. I carried it around my neck on a lanyard or in the pocket of my camera bag when I was out and about. A supplied light adhesive patch kept it on my dashboard while driving. The unit also comes with both mains (AC) and car chargers.

That’s the good. The bad is that while GiSTEQ says CD110BT will work on PCs and Macs, mine didn’t out of the box. It required a firmware update to work with a Mac, and the firmware updater only works on PCs and didn’t like Windows XP running on Parallels virtualisation software. Fortunately, my friend Andy Carvin at NPR gave me five minutes on his PC to update the firmware, but even after that, I had difficulty getting the device to consistently download data. GiSTEQ has since released a new update that they say fixes this. I downloaded some GPS logs tonight without a hitch.

I’d like to try the Amod AGL3080 (review in Wired), which is touted as a driverless geo-tagger. It simply mounts as an external drive on Mac or PC, and all you need to do is copy the data from it. It uses a highly accurate SiRF III chipset. Unlike the GiSTEQ which is charged via the USB cable, the Amod runs on three AAA batteries. Kevin Jaako has a thorough review of it on his blog.

The software that comes with the GiSTEQ promises a lot and delivers most of it without too much fuss. It’s actually rebranded software from JetPhoto, and as the company says on its site, you don’t actually need a specialised geo-tagger. There are several Garmin or Magellan GPS units that will work with it. The software also works quite nicely with the N82, instantly recognising that the photos already have geo-data embedded in the files. If the geo-data is off, the software has a nice interface to relocate and update the geo-data. It also has a built-in Flickr uploader, although it could be a bit more intuitive and work more seamlessly with Flickr title and description fields.
But I didn’t just geo-tag my photos. I also geo-tagged my tweets using Twibble, a geo-aware Twitter app Nokia S60 phones. Twibble integrates seamlessly with the GPS on the N82. It also allows you to upload pictures you’ve taken with the phone directly to TwitPic. We just used this all to great effect for Guardian Travel’s first TwitTrip with Benji Lanyado. It is pretty heavy on the battery, but I had a power inverter in the car so everything was fully charged all the time. It was also a bonus to have Nokia and Google Maps on the phone for navigation.
I also geo-tagged all of my blog posts. I either took the geo-data from a Tweet or a photo, or if I didn’t have any geo-data handy, I used sites like Geo-tag.de or Tinygeocoder.com to generate geo-data from an address.
Visualising the trip
Thanks to a quick bit of python scripting by Guardian colleague Simon Willison, I have a KML file for all of the 2059 photos that I took over the more than 4000 miles of the trip. One of the reasons that I wanted to geo-tag pictures, posts and tweets was that while I know most of these towns, I wanted to give a global audience a sense of place.

But apart from easily visualising the trip, why all the fuss to do this? Adding geo-data to content is one of those fundamental enabling technological steps. It opens up a world of possibilities for your content. By geo-tagging your content, it allows users to subscribe to content based on location. Geo-tag your movie and restaurant reviews, and you can start leveraging emerging location-based services on mobile phones. With Google Maps on mobile and other mapping services, news organisations could provide real-time location based information. Geo-data allows users to navigate your content by location instead of more traditional navigation methods.
Some companies are already dipping their toes into geo-data. Associated Press stories hosted on Google News have a small inset Google Map inset based on the location information in the dateline. New York Times stories appear on Google Earth. But datelines are imprecise because they are city-based, but when you pull up more accurate data you can do much more. You can see the possibilities of mapped information on Everyblock.com.
But to get from most news sites to Everyblock, you’ve got to put in the foundational work both on the technical side and the journalistic workflow. Having said that, it’s not rocket science. It might seem a lot of work up front, but once the work is done, geo-data provides many opportunities, some of which could provide new revenue streams.