NICAR day one has been a whirlwind, but I’ve already learned a ton and had quite a bit of fun. Here are the quick highlights from the day:
NewsCamp: Investigating text in the wild
I was super excited for this panel. Picture having to analyze hundreds of government documents. And, we’re not talking about a tab-delimited table; just raw text.
Sarah Cohen from Duke University walked through some fantastic examples of how newsrooms/companies are using new technologies to analyze raw texts and some of the limitations out there, such as accurate sentiment analysis (sarcasm is hard for a computer to recognize).
I could see us potentially doing similar raw text analysis of calendars for state legislators in Delaware. It would be cool (depending on the thoroughness of the calendar) to be able to tell which lobbyists/deal-makers were meeting most with particular legislators, especially if the dates correspond with big votes.
JamSession: The Natural Language Tool Kit
I was a bit disappointed that (a) I didn’t get a better seat closer to the projector and (b) I wasn’t able to stay for this whole session (had to speak on a different panel).
The JamSession was an introduction to the Natural Language Tool Kit (NLTK), which is a fascinating Python Library for analyzing raw text (see above section). I’ve played around with the demos a little bit, but I know that this tool has much more to offer.
What I was able to gather from the portion I attended, though, was pretty informative and reinforced my thought that NLTK should be a tool I become closely acquainted with.
That darn Chris Spurlock made us walk halfway across the city… worth it, though.
Locating the story: The latest in mapping
One of my former professors, David Herzog, lead this panel with Ben Welsh of the LA Times. I was already pretty familiar with most of the technical aspects (spatial joins, etc), but the mapping examples were pretty inspiring. Some of the story ideas that came out of the session were:
- Acquiring and mapping health department records on respiratory/asthma issues to see which neighborhoods produce more problems.
- Foreclosures. One of the featured maps was fascinating because it showed that banks were foreclosing one home and buying the neighbor’s tax lein. With Wilmington being such a big banking/credit card city, there could be a lot of potential in this data.
- Racial Disparity. Stories such as, which neighborhoods have their potholes filled first and what are the demographics like.
Welsh also touched on an issue The News Journal might be facing soon: Google now charging for its mapping service. I’m going to have to study this a bit more. We might be switching to one of ESRI’s base-maps for our projections.
Turning your stories into a tablet/phone app
Look for an additional post on this subject later tonight. There was a lot of information, and I’d love to establish a separate venue for brainstorming ideas on how The News Journal benefit from creating e-books, HTML5 apps, etc.
Car on a shoestring
This was the panel I spoke on with two other, very talented, journalists.
Check out our presentation here.