A little set up is required for this one so bear with me…
The business section of The Atlantic shares my reaction to reading the most recent unemployment numbers. Their article, entitled “Unemployment Rate Fall to 9.7%, But 20,000 Jobs Were Lost?”, elaborate by saying…
The national unemployment rate declined in January to the seasonally adjusted rate of 9.7% from 10.0% in December, according to the Bureau of Labor Statistics. The rate beat consensus expectations of remaining unchanged at 10.0%. But even though the national rate declined, 20,000 more jobs were lost. That was worse than consensus expectations, where economist expected a gain of 15,000 jobs.
So why did this happen? The answer is that the numbers were “seasonally adjusted” for January…
Remember, the reported number is seasonally adjusted. It says that in January, we had 14.8 million unemployed Americans, which is about a half million fewer than the 15.3 million unemployed in December. Sounds great, right?
Maybe, maybe not. If you don't adjust for seasonality that changes. A lot. Then, January had 16.1 million unemployed, while December had 14.7 million -- an increase of 1.4 million jobless Americans.
In other words December is a month of Holidays and more people were hired specifically for that month. Those people were then “laid off” in January but the unemployment rate doesn’t count them which is why the rate went down even though the number of jobless went up.
At this point you’re probably wondering why I’m writing about this on a tech blog. Because the journey to getting that explanation led me through the Government’s data repository site Data.gov.
For those who haven’t read my posts in the past I’m not a fan of Data.Gov. I applaud the idea but the effort itself is so badly done that it’s next to useless. The perfect example of this is the fact that job statistics aren’t published on the site.
In fact, I couldn’t find much of anything useful. The closest I found was a report entitled “Employment, Hours, and Earnings from the Current Employment Statistics”. The description said…
Each month the Current Employment Statistics (CES) program surveys about 150,000 businesses and government agencies, representing approximately 390,000 individual worksites, in order to provide detailed industry data on employment, hours, and earnings of workers on nonfarm payrolls.
So I clicked on the little CSV icon…
and came to this…
Yes, apparently I need Java to download a simple CSV file (CSV stands for Comma Separated Value and is basically a text file with values separated by commas) . So I went along with it and that brought me to this…
Which finally led me to this (In a pop-up if you can believe that)
A few things to note here…
1. The whole point of Data.gov is to make the data sets available to programmers. But this has so many loops to jump through (Java, Popups, redirects, etc…) that you’d need a sophisticated crawler to actually get the info (which defeats the whole point)
2. Data.gov goes on about how many data sets they have available and what a success the site is because of that. But if they’re just reposting links to department websites that existed before the site was put up than is it really successful?
3. In the end Data.gov didn’t have the employment data that’s distributed to every major news organization. So how effective is this initiative if the Dept. of Labor isn’t even putting all it’s most recent data on the site?
So again I go back to the point I try to make regularly here which is that doing anything (no matter how ineffective) is not better than doing nothing at all. Every time someone praises the success of Data.gov what I see is the cover up of a system that doesn’t really work at all.