On Weather

In yesterday's New York Times was an article that's completely reflects the purpose of this blog. They explain that there are less murders in NYC on rainy days than on sunny days.

Being that they did the analysis for me (thanks, guys!), let's take a look on how an article like this gets written. It's possible the intrepid Times reporters started with a regular review of historical daily murder figures, matched them with weather patterns, and came up with a correlation between days of rain and homicides.

Being that the recent weather in New York City had been so damp in the month of June, it's possible that the reporters (or an editor, or an unrelated analyst) saw the decrease in year-on-year murders from last year's really hot and dry June, brought it to the attention of the right people, and from there the reporters decided to investigate further.

I'm going to venture a guess and go with that second scenario.

Most detailed data analysis is founded in brainstorms. "Hey, what do you say about this?" or "This might be a pattern." Investigations do not always yield measurable results that can be neatly packaged into an article. But this one did.

The best part about this article is that it doesn't just provide a clear explanation proving a thesis. It gives lots of possibilities as to why the analysis might turn up the data they are presenting. This is a big difference. Data doesn't lie, but why it falls into patterns is not always so cut and dry.