Monday, July 02, 2012

Data, information, knowledge - The Great Pittsburgh Population Caper Continues

Ok Yunzerwonks... switch out of counting in binary and check your significant digits.  The story of the great Pittsburgh population caper is getting stranger.  You will need to follow through to the end. When you finish consider that if it is so hard to jump the transom from data to inkling for such a simple story as this, consider the implications for some of the much more complicated stories we all obsess on routinely.

Anyway, see PG today for some confirmation on what some census data released last week really meant for Pittsburgh:  Census data shows unusual rise for Pittsburgh. One clear point from that is that the city's estimate here includes a rise of approximately 1,200 in the city's Group Quarters population (namely college students in dorms, prisoners, nursing homes and related) and that is the biggest factor in why these numbers look as they do.  . 

I won't repeat myself, but you can read a couple of the posts from Thursday and Friday to go through the big picture confusion over this data release that applies to most of the nation.  To put a point on the Pittsburgh numbers though here are some specific points to ponder.  I've deconstructed the data from this particular release to highlight the population estimates for Allegheny County and the City of Pittsburgh and how the change in group quarters population impacted the observed changes.


So take out the gain in group quarters population and the only nominal increase in population is coming from the assumption that the city's population grew at the same rate as the county's population.  No new information or analysis behind that assumption, it's just the methodology they used for this particular data because they didn't have the data they normally used for this particular annual data product.   If you don't believe that.. take the Pittsburgh net growth of 670 as a proportion of the county's net growth of 2,529.  Works out to 0.264.  Now take the city's population (305,704) as a proportion of the county's population (1223348).  Works out to 0.250.  So the new population is an artifact of methodology, and not reflective of any 2011 data for the City proper.

Only the beginning.

Beyond group quarters, maybe the city's population did grow, or maybe it shrunk.. the point is there is no information in this particular data to lead us to believe one way or the other.   

So to get at least from data to information you have to look into where the data comes from.  That is always true.  Just reporting a number, any number.. even a 'Census' number takes more work than that.  Now go back and read the gratuitous Pittsburgh references in the LA Times coverage of this exact same data:  U.S. population in cities growing faster than in suburbs

To be clear.  There is nothing that "doesn't count" about Group Quarters populations.  If it was prisoners (aka the increase at the Allegheny County Jail Downtown in the 1990's) then the implications for the city's future growth may be minimal.  But college students in dorms or those in nursing homes actually have an economic impact in the city. The demographic implications are pretty different, but hold that thought.  Students in dorms have had a big part of the changing demographics of the city of Pittsburgh in recent years so there is nothing surprising about this.  Realize that students in dorms are only a part of the total student population, likely only a small part of the graduate student population which has seen marked growth at some local institutions.

So in the case here.. the new Group Quarters population was all from an increase in college students in dorms.  No increase in any prison population, nor any nursing home or military barracks (not that any exist which I am aware of) were part of this data.  So we have a story of a big new jump in the student population in the city eh? 

Not really it turns out. There is a longer term story here for sure, but the story in the latest data is evaporating quickly.   So it looks like most of the increase in the dormitory population may be coming from a dorm downtown that was... well..  missed in the 2010 census.  Looks like it was a dorm that opened in 2008 and somehow was overlooked and we are likely talking close to 800 folks which were put into the 2011 estimates data.   That would make a lot of sense since if you looked at the city's population change, it all happened in the 4 month period between the April 1, 2010 reference data for the 2010 Decennial census and the July 1, 2010 reference for the annual population estimates. That in itself didn't pass the smell test, which is in fact what set me off on this little mystery.

Lots of implications to that.  For one is that the Downtown population numbers may be off, as would be the city of Pittsburgh and Allegheny County.  It also means that if you are looking at this latest release of population estimates for 2011 and imputing a growth rate over just a year and extrapolating forward.. then you may be extrapolating a lot more than is warranted.  It is in a sense a half empty-half full thing (where is ADB?).   There may be more people living in the city than was thought, but then the 'growth' in this data is a lot slower than it even appears nominally. 

Downtown population is always a bit tricky because of past Census Bureau errors.  In 2000 the Allegheny County Jail was misplaced in its geography in the census.  When the new jail was built it moved from Dowtown to the Uptown neighborhood technically.  That point was missed and the jail's population was all placed Downtown at its former location.  That error makes the decade over decade comparison difficult.  I once tried to do my own calculation of Downtown's residential population for the record to account for those errors.. but I am thinking now this new omission will require that work to be redone. 

1 Comments:

Blogger Scollon Taylor said...

Middle and high school students in Korea voted miss A’s Suzy and IU as two female celebrities who seem to be bad at math.

On August 17th, K-Stem, an institute that conducts a scholastic ability test in mathematics, announced the result of the survey asking 220 students to vote for the celebrities who seem to be good/bad at math.

For the category asking student to vote for celebrities who seem to be good at math, actress Kim Tae Hee ranked first with 43.4% of the votes, and Lee Seung Gi followed with 30.9% of the votes. According to the students, “Lee Seung Gi looks smart and intelligent” and “Kim Tae Hee went to Seoul National University, she is smart for real.”
บอลวันนี้ Casino With Bonusreverse mortgage los angelesOyunlarmama liciousbest franchise opportunitySBOBEThealthy food for a dietfree pornNew York EscortsVoucher Codesjobsonly online shopInterior Designers MiamiVoucherCodesNetGet Rid Of Timeshare บอลวันนี้ Casino With BonusPromotion Codesfake watch

On the other hand, for celebrities who seem to be bad at math, Boom ranked first with 45.15% of votes, followed by female idols Suzy and IU who received 29.1% and 22.9% of votes respectively.

Friday, August 17, 2012 10:45:00 AM  

Post a Comment

Links to this post:

Create a Link

<< Home