MetaCarta and H1N1

I had John Frank, founder and CTO at MetaCarta on the phone on another matter (see this post) and the discussion turned to mapping H1N1.

MetaCarta held a webinar this week for intel type folks on just that topic. Frank shared some of the concerns I did about maps and data sources. He noted that the number of locations mentioned related to the flu total about 71,000 and many, many of those are not making it onto the maps that officials and the public are viewing. For example, there are many cases in Asia/Pacific that are not "registering on maps" in part because of where those making the maps are looking for information.

That led me to ask about how data about H1N1, for example, can be vetted for accuracy. The traditional way for the intelligence community (and journalists too) is confirm the same information from independent sources. The research along, aside from the analysis, as you might expect is a time consuming process! (I have in my head a comment from my friends who were National Geographic interns that nothing goes on the map until its confirmed by at least three independent sources.)

So, how do you do that in this type of situation where say a confirmed case is reported in a small town in New Zealand? That, Frank suggests, is where a tool like MetaCarta can come in. It can "do the legwork" of finding mentions of say "flu" in unstructured text (news reports among other things) that map to that location or ones nearby. Now, that in itself doesn’t confirm anything, but gives the analyst a starting point to find, validate and confirm independence of the sources.

More to Frank’s point of the missing data points in some geographies - MetaCarta tools may help identify geographies that are not yet "on the map" of the disease that perhaps should be.