These are non-directed dyad-year data for the minimum distance between states in the Correlates of War state system from
1946 to 2015. The data are generated from the cshapes package.
cow_mindistA data frame with 817053 observations on the following 4 variables.
ccode1the Correlates of War state system code for the first state
ccode2the Correlates of War state system code for the second state
yearthe year
mindistthe minimum distance between states on Dec. 31 of the year, in kilometers
The data are generated from the cshapes package. The package authors purport that the data
are generated to be compatible with Correlates of War system codes, but a review I did several years ago for
an unrelated project (published in 2017 in Conflict Management & Peace Science, which you should cite for
all your articles if you're reading this) suggested the output does not seem to perfectly meet that billing. These
included oddball cases like Zanzibar, United Arab Republic, Comoros, East Germany, and a few others. I pre-process
these as outlined in the associated file in the data-raw directory on the project's Github.
Data are automatically generated (by default) as directed dyad-years. I elect to make them non-directed for space considerations. Making non-directed dyad-year data into directed dyad-year data isn't too difficult in R. It just looks weird to see the code that does it.
Most of the data I prove elsewhere in this package are to be understood as the data as they were at the *start* of
the year. This is how I process, for example, the capitals data as they get merged in the add_capital_distance()
function. However, the script that generates these data are set at Dec. 31 of the year and not Jan. 1. I do this for concerns
of maximizing data coverage. If you wanted the same effect, just lag the data a year.
Weidmann, Nils B. and Kristian Skrede Gleditsch. 2010. "Mapping and Measuring Country Shapes: The cshapes Package." The R Journal 2(1): 18-24