Thursday, November 05, 2015

Computation and Journalism Part 2

The theory blogosphere is atwitter with an announcement of a talk next Tuesday at Chicago by László Babai Graph Isomorphism in Quasipolynomial Time. Luca will blog from the scene and we will have a full report as details emerge. 

More from the Comp and Journalism conference

0) (ADDED WRITE BEFORE POSTING) The conference was about how journalists do or should use technology. Part of this is how news gets to people. Twitter and Blogs are one method, as the above note about Babai's talk shows.

1) There was a panel discussion on context When one reads an article it depends on where one is coming from. Here is a strange thing that technology has allowed journalists to do:  An article can have slight differences depending on where its being read. For example if someone from New York is reading it may refer to Broadway, whereas if someone in LA is reading it it might refer to Hollywood. The tech also exists to have the reader (online) have a choice- so you can read it as if you are in place X. This scares me a bit- what if an article is can be tinkered with to appeal to lefthanded latino lesbians who work on the local
Lovasz Lemma.  Isn't our country fragmented enough?  STILL, the point that context matters came through nicely.

2) There was a panel of three ACTUAL JOURNALISTS (I think that all of the other speakers were academics) who used BIG DATA or MACHINE LEARNING or FILL IN BUZZWORD to write articles.  Propublica, a nonprofit newspaper (aren't they all?) did an article about rating doctors (here ) that seems to be very well checked (example- if a patient has to come back again because of a complication, is it the surgeons fault? hard to answer).  However, some doctors (those that got low ratings?) complained that there article was not peer reviewed. This brings up a question--- journalists run things by fact checkers and proofreaders, but do not have peer-review. Academics do. Which is the better system?  Another journalist  did a study of the Supreme court and found that there is a small set of lawyers  that are well connected (one used to play darts with Alito) who are involved in  over half of the cases the court sees. What interested me is--- what if you have  an idea for an article and you do ALL of the number crunching, bring in ML people  to help and find out in the end OH, the Supreme court actually is doing a fair job or OH, Doctors are pretty good. No story! They said this happens about 1/3 of the time where either you don't have enough data for a story or the story is not really a story.

3) There was a paper about YELP ratings of doctors. It seems that the actual
doctor-stuff is rarely mentioned, usually it was how long the wait time was,
how boring the magazines in the office were, etc.  Also, most doctors got
either 1 star or 5 stars. One thing they didn't mention but I will- Walter Palmer, the Dentist who killed  Lion in Africa, got lots of very negative YELP reviews from people who were never patients of his. So YELP ratings can be skewed by things independent of what is supposed to be measured. That is, of course, and extreme case.

4) Key Note by Chris Wiggins who is an academic who, on his Sabbatical, worked for the NYT on issues of How many copies of the paper should we print? and other  Operations Research type questions.   Print  subscriptions still out number online subscriptions, the online is doing pretty well and bringing in money, but they still NEED to find a better business model.I am reminded that when Jeff Bezos bought the Wash Post Stephen Colbert said there are more people buying newspapers than there are people buying newspapers.
(My spellchecker things online is not a word. Everyone else does.)

5) There was a panel on Education. From there point of view very pessimistic---
most schools don't offer courses in how to use Technology in journalism, no real
books, no agreement on what's  important.  And I suspect they will have a hard time updating courses once they have them. I wonder if the early days of Comp Science were like that.

6) The issue of jobs in journalism going away was not quite ignored but also not
quite confronted either. Some people told me that Journalism schools are not being honest with their students about the dismal job market. Then again, they are journalism students- they should investigate!

7) A journalism student told me of a case going to the supreme court that may be
very important for privacy: Spokeo vs Robins: here

8) UPSHOT- I am glad I went, I learned some things outside of what I usually think about.  Unlike the RATLOCC (Ramsey Theory and Logic) I was able to tell my mom what I learned.  I didn't bring my laptop nor had TV access I was able to work on some math and had a minor breakthrough. But that,s for a later blog
(My spellcheck thinks blog is not a word. It also thinks spellcheck is not a word.)

No comments:

Post a Comment