Pittsburgh GDP up 3.3% (sort of)
So for a reference number if you want to know how big the Pittsburgh economy is, the current dollar estimate for metropolitan GDP is $138.9 billion annually. GDP is, broadly speaking, intended to be a measure of value added production. A good number to keep in mind when comparing the size of things you read about in the news.
But I need to look at something closer here. The real GDP estimate for Pittsburgh shows a 3.3% jump, but the current dollar increase between 2014 and 2015 is only showing a 0.9% increase. Inflation is low, but typically even with low price inflation current dollar increases appear larger than the inflation adjusted amount in almost all time series. Looking at this data, for most all regions current GDP growth is higher than real GDP growth, but for Pittsburgh it isn't, by a palpable difference. So I calculated the ratio of the current dollar GDP to inflation-adjusted GDP across all metropolitan areas in the BEA data. The ratio of the two is different across all MSAs, but within a narrow range of around 1.12. For Pittsburgh, the ratio is much lower at 1.06, one of the lowest ratios among all MSAs that I calculated.
So now we get all wonky, but this is interesting. Basically if the data is correct, it is saying there has been an abnormal deflation in prices in Pittsburgh based on the mix of goods consumed here. The methodology used by the BEA does not assume different regional prices per se, but it does de facto calculate different regional price deflators based on different mix of goods produced here. So if we produced more of some good that had abnormal price changes, that results in regional differences in deflators and the result observed. If I had to speculate, I think the strange result for Pittsburgh (much lower current dollar GDP growth compared to 'real' GDP growth) is coming from price deflation in energy markets. The price collapse in coal and natural gas over the time period for this data would give this result for regions that have concentrated output in those products. Makes for some different interpretations of the data for surer.
So yeah, methodology matters, no matter how easy it is to quote a headline number.