search

csarven / worldbank-linkeddata

World Bank Linked Data
http://worldbank.270a.info/
Other
9 stars 9 forks source link

Update to latest Barro-Lee population model? #1

Closed samuelleach closed 10 years ago

samuelleach commented 10 years ago

I've been using your linked data - the Barro-Lee population model - from the World Bank site, using the R package 'WDI' (http://cran.r-project.org/web/packages/WDI/index.html).

For instance I'm using the indicators BAR.POP.3034, BAR.POP.3539, etc to forecast the population of the Barro-Lee like age segments in the middle east.

I was lead to your data via this page http://worldbank.270a.info/classification/indicator/BAR.POP.3034.html)

What I'm finding is that when I total up the Barro-Lee age segments (ages 25+) along with World bank indictors (SP.POP.AG03.TO.UN etc) for ages 0-25, then I find that my totals don't match very well the 2010 total populations as shown in this type of explorer:

https://www.google.co.uk/publicdata/explore?ds=d5bncppjof8f9_&met_y=sp_pop_totl&hl=en&dl=en&idim=country:EGY:SAU#!ctype=l&strail=false&bcs=d&nselm=h&met_y=sp_pop_totl&scale_y=lin&ind_y=false&rdim=region&idim=country:BHR:JOR:KWT:ARE:QAT&ifdim=region&hl=en_US&dl=en&ind=false

The discrepancy is greatest for countries like Qatar, UAE and Bahrain that have seen accelerated total population growth in the last year.

e.g. Barro-Lee give 4.8M total population in 2010, while Google (based on Worldbank totals) give 8.4M.

My question is, is it possible that the Barro-Lee data provided by the World bank, and your linked data, is not using the latest Barro-Lee data, which came out in April 2013 (according to their website)?

csarven commented 10 years ago

@samuelleach I don't work on the cran or GPDE, so I can't speak for them as to why their values don't match with worldbank.270a.info's. What was your SPARQL query to gather the data?

I will synchronize with World Bank's data sometime soon. IIRC, the last update I did was about 4 months ago. If there is a discrepancy between the values from World Bank's API and my Linked Data, please let me know. I can only address that.

samuelleach commented 10 years ago

I have been using the R stats package WDI to hit the World Bank API. Here is a typical R call for three countries whose total population (when summed over SP.POP.AG.. and BAR.POP..) seems to be off:

DF <- WDI(country=c("AE", "BH", "QA"), indicator=c("SP.POP.AG00.TO.UN", "SP.POP.AG01.TO.UN", "SP.POP.AG02.TO.UN", "SP.POP.AG03.TO.UN", "SP.POP.AG04.TO.UN", "SP.POP.AG05.TO.UN", "SP.POP.AG06.TO.UN", "SP.POP.AG07.TO.UN", "SP.POP.AG08.TO.UN", "SP.POP.AG09.TO.UN", "SP.POP.AG10.TO.UN", "SP.POP.AG11.TO.UN", "SP.POP.AG12.TO.UN", "SP.POP.AG13.TO.UN", "SP.POP.AG14.TO.UN", "SP.POP.AG15.TO.UN", "SP.POP.AG16.TO.UN", "SP.POP.AG17.TO.UN", "SP.POP.AG18.TO.UN", "SP.POP.AG19.TO.UN", "SP.POP.AG20.TO.UN", "SP.POP.AG21.TO.UN", "SP.POP.AG22.TO.UN", "SP.POP.AG23.TO.UN", "SP.POP.AG24.TO.UN", "SP.POP.AG25.TO.UN", "BAR.POP.2529", "BAR.POP.3034", "BAR.POP.3539", "BAR.POP.4044", "BAR.POP.4549", "BAR.POP.5054", "BAR.POP.5559", "BAR.POP.6064", "BAR.POP.6569", "BAR.POP.7074", "BAR.POP.75UP" ), start=1980, end=2015)

BTW, The WDI package source code shows the actual API that is hit: https://github.com/vincentarelbundock/WDI/blob/master/R/WDI.R Is it the same as yours?

Overall I am still wondering if the World Bank has updated to the latest Barro-Lee model, and I have posted a question on their UserVoice forum.

csarven commented 10 years ago

This repository is about the realization of World Bank's data published as Linked Data: http://worldbank.270a.info/ . See also: http://270a.info/

It has nothing to do with the R package you are looking at or Google Public Data Explorer and so on. The URL of this repository is clearly different than the package you are dealing with. I suggest addressing your question at the appropriate repositories or mailing lists.