Wednesday, August 31, 2011

Sharing a data dashboard

When you are the City of New York, sharing data -- a good thing -- inevitably means that someone will criticize you. The city shares its performance data through something called the Citywide Performance Rating; it's available here. It's a successor to the old Mayor's Management Reports and collects and reports on city agency performance in eight major areas (called "Themes" on the city's website): Citywide Administration, Community Services, Economic Development and Business Affairs, Education, Infrastructure, Legal Affairs, Public Safety, and Social Services. The data are updated monthly.

Here's a screenshot of the Citywide Performance Data main page:


In future posts I'll look at each area in more detail. For the moment, here's a link to an unusually thoughtful article in the NY Times about using data - and, perhaps just as much, about the vicissitudes of making one's data dashboard public.

Monday, August 29, 2011

On hurricanes and predictions

Updated 30 August
 
It's become fashionable in the last 24 hours, now that Irene has come and gone from New York City, to say that officials overreacted and that NYC's mayoral administration, in particular, had to live down its dismal performance during last year's snowstorm. Here's how a local street looked for several days after that storm:


But I don't think Mayor Bloomberg overreacted to the threat Hurricane Irene posed (for one thing, it's too easy to say that after the fact; for another, there were big trees down in my neighborhood and flooding not so far away). Once again, interpreting the information we had gives a lesson in how important it is to understand the basis for the news.

In Irene's case, the issue was the likelihood of a storm following a particular track, and its intensity when it arrived. After the fact, I think that Irene changed course slightly, hitting land to the west of NYC, not the east, and weakened in intensity. It clearly caused a lot of damage - several people have been killed, there has been extensive flooding, and lots of property and tree damage. The damage was lighter here. In advance that was not so certain - and the storm did hit at high tide, with high winds and lots of rain. Sites like weather.com provide useful quick predictions, and lots of scary graphics. But they're interpreting readily available sources of information, and you can look at them yourself and  make your own judgments.

In addition to the map I posted Friday from the NY Times, a site I found very useful was the National Weather Service National Hurricane Center site. Click through it to the Irene Graphics Archive and you can see that the NWS' prediction of Irene's path was fairly accurate. The cone moves, but here's a screen shot of the five day cone/warning from one of the Wednesday briefings:


(The New York Times carried a story yesterday about how much better meteorologists are getting at predicting storm tracks, and how hard it can be to predict the intensity of a storm at any particular time. And BTW, I'm impressed with the Times for taking down its paywall for the duration.)
UPDATE: Here's a link to NPR's story aired Monday, 29 August, on the same subject. (Water temperature, land topography, air temperature to name a few elements.)

There was a lot of talk about Nate Silver's tweet stating that a Category 4 hurricane making landfall in New York City would cause damage roughly comparable to the damage caused in Japan by the earthquake/tsunami combination. If you read the full post, (always worth doing with Silver) you'll see all of Silver's qualifiers - starting with the crucial fact that a Category 4 storm has never hit the Northeast US. And in any case, there was a great deal of concern that Irene might come ashore here as a Category 3 storm (it wound up a Category 1 here).

After scooting around on a lot of sites, I decided that I should be a little, but not a lot, worried. So how did I prepare? I found Melissa Clouthier's list to be the most helpful. I knew I had bleach and paper goods, and I figured guns and ammo would not be an issue here in NYC. Mostly what I did was buy fresh fruit and fill water bottles (wondering how we had accumulated so many) and check the candle supply.

As usual: forecasters are basing their predictions on statistical models. Weather reporters are not always trained as meteorologists. Excitement means big news. So - take what the reporters are telling you with a grain of salt, and draw your own conclusions. And then turn on the Weather Channel to watch the waves crashing.

Friday, August 26, 2011

Good Hurricane Irene Graphic

If you're tracking Hurricane Irene, as I am, the NY Times has a great graphic of the storm's projected path. Here's a screen shot, as of Friday morning:

But click on the link, the map updates regularly (also on the Times site the little hurricane eye keeps spinning.)

Thursday, August 25, 2011

Math education in the Times Op-Ed

There's an interesting Op-Ed piece by Sol Garfunkel and David Mumford in today's New York Times, arguing that high school math education should be contextual or applied, rather than pure. There's something to be said for their point of view, though I do disagree with their example about Latin. What do you think? Discuss in the comments.

Thursday, August 18, 2011

On interpreting data

Yesterday's New York Times carried a story reminding me again that when working with numbers, you can't look at them just once. The story, titled, "On Economy, Raw Data Gets a Grain of Salt," by Binyamin Appelbaum, explains how the federal government's Bureau of Economic Analysis changed its estimate of the growth rate of the economy for the first quarter of 2011. Instead of growing at a rate of 1.8%, the government now reports that the economy grew only the at the much smaller rate of 0.4% during that quarter.

The reason? The Bureau of Economic Analysis changed the value of vehicles awaiting sales at dealerships. In the initial reports, these numbers had been projections, not actual counts; those come later, and are starting to be available now. The change in that one number thus had a profound impact on our understanding of what happened earlier this year.

A couple of lessons:

   * In this kind of analysis, you're using certain numbers to signal or represent what's happening in the larger world (in this case, the economy). If you do use an index number like this, you need to revisit it periodically, to be sure it's representing the information you think it should.

   * It's important to understand when a number is solid (or as solid as possible) and when it is an estimate. As Appelbaum says in the story, "[P]oliticians and investors are placing a great deal of weight on a crude and rough estimate that has never been particularly reliable." Back in January, I wrote a blog post about a very useful book called Proofiness: The Dark Arts of Mathematical Deception, about the habit of succumbing to the authority of numbers. You can read the post here, but better yet, read the book!

Economic numbers may be particularly opaque to interpret, but the story has some relevance to anyone using numbers to manage and understand their business, whether it's for-profit or not-for-profit. Using numbers is an iterative process; you need to examine your numbers and think about what you're doing over and over again.

I'm back briefly between trips, and wanted to get this post up. More to come next week.

Popular Posts