oc,wmay

wmay / oc Goto Github PK

View Code? Open in Web Editor NEW

oc — Optimal Classification Roll Call Analysis Software

C 2.08% Fortran 85.16% R 12.75%

oc's People

Contributors

Stargazers

oc's Issues

Estimated polarity of each bill not available to user

The oc object does not store its estimate of the polarity of each bill, i.e., whether OC predicts a Yay or a Nay for being on the left of that bill's cutpoint. (see reprex in #2). Would be great to have from a user perspective.

cc @JeffreyBLewis

plotting raises an error

In R >=4.2, plot.OCobject raises an error:

Error in axis(2, at = yt, ...) : 
  formal argument "at" matched by multiple actual arguments

Is the rollcall output ordered the same way as vote.names

It would be reassuring to have the rownames of the $rollcall output be named by the rollcall vote IDs as identified in the rollcall object. I presume it is ordered the same way as the columns of the input matrix (?), but better if it's tracked.

This is done in the legislator output:

oc/R/oc.R

Line 208 in 99c9a7b

rownames(legislators)<-rownames(rcObject$votes)

library(oc)

data(sen90)

# input data has vote IDs
head(colnames(sen90$votes))
#> [1] "Vote 1" "Vote 2" "Vote 3" "Vote 4" "Vote 5" "Vote 6"

# Run OC
sen90.1d <- oc(sen90, dims = 1, polarity = c(7))

# But output data has no rollcall IDs
head(sen90.1d$rollcall)
#>      correctYea wrongYea wrongNay correctNay        PRE normVector1D midpoints
#> [1,]         38       13        1         48 0.64102564            0     55.50
#> [2,]         47        1       13         38 0.64102564            0     55.50
#> [3,]         52        6        2         40 0.82608696            0     59.50
#> [4,]         50       22        4          6 0.07142857            0     14.75
#> [5,]         68        5        2          9 0.50000000            0     14.75
#> [6,]          2        0       15         75 0.11764706            0      3.50