Home

Welcome to the qogdata wiki!

Purpose

The xtdata package aims at providing some helper functions to users working with Quality of Government (QOG) datasets, especially with regards to merging different country-year cross-sectional time series. The package allows the user to download data from the QOG data server and from several sources, as well as to merge, plot or subset it according to the specifications of the xtdata data frame attribute.

Readings

• Jerven, Morten (2013), Poor Numbers. Cornell University Press.
• Rothstein, Bo (2011) The Quality of Government. University of Chicago Press.
• QOG Reading Recommendations (time series methodological articles)

Technical

• Here is a complete description of the xtdata specification attribute for data frames, with planned improvements.
• Here are the Stata do-file and R scripts to generate the QOG variable indexes and demo dataset bundled in the package.

TODO

• done, 0.1.1 fix dependencies
• done, 0.1.1 let xt discover multiple panel data types
• done, 0.1.1 let xtcountry discover additional identifiers
• done, 0.1.1 let xtmap accept other formats than iso3n
• done, 0.1.1 xtsubset to subset panel data while preserving xtdata attribute
• done, 0.1.1 code xttse and xtdecay as in time-since-event and decay functions
• done, 0.1.2, make xtlag and xtlead robust to k larger than the series
• done, 0.1.3 enhance xtplot to work on subsets and with tons of geoms
• done, 0.1.4 fix xtmap when using facets and/or quantized time periods
• done, 0.1.5 xtdes Stata equivalent
• planned, 0.1.6 code xtsum xttab Stata equivalent
• planned, 0.1.7 code xttab Stata equivalent
• planned, 0.1.8 idea: recode better indexes
  • get Stata variable labels from R after importing with foreign
  • add value label column and label and label table
  • make qogfind search the first and qoglabel search the second
• planned, 0.1.9 code qoglabel for getting QOG value labels
• planned, 0.2.0 code get_eurostat to get Eurostat national and regional data
• planned, 0.3 let xtmap plot maps of the US based on state and state.fips
• planned, 0.3 let xtmap plot maps of france (subset NUTS-2 region to French ones)
• planned, 0.3 code xtregion to handle NUTS regional codes (http://christophergandrud.github.io/DataCombine/)
• unplanned CSTS modelling tools:
  • idea: code something like dataRep in Hmisc
  • idea: support simcf (Zelig simulations for TSCS models)
  • xtreg etc. Stata equivalents?
• unplanned data tools:
  • idea: add export to DataPackage format, using RJSONIO
  • idea: work with RSDMX and XML to read SDMX formats
• turn xtdata into a proper class