Using wru offline

[emcghee73]

I have data on a server without access to the internet, and I'm not allowed to remove the data from that server. Can I use wru without internet access? I tried pre-downloading the census data and using it in the predict_race function, like so:

census20 <- readRDS("wru_census_2020.rds") #loading the pre-downloaded census data
d <- predict_race(filter(d, !is.na(surname)), census.geo="tract", census.data=census20,
year="2020", age=F, sex=F, party="party",
names.to.use="surname, first, middle", model="fBISG")

but I got this error message:

Using predict_race to obtain initial race prediction priors with BISG model
Proceeding with first, last, and middle name predictions...
There was an error retrieving dataCannot access release data for repo "kosukeimai/wru".
Error in gzfile(file, "rb") : cannot open the connection
In addition: Warning message:
In gzfile(file, "rb") :
cannot open compressed file 'C:\Users\mcghee\AppData\Local\Temp\4\RtmpYpAfz4/wru-data-first_c.rds', probable reason 'No such file or directory'

If it's not downloading the census data, what is it downloading? Can I do without it, or pre-download that, too?

[1beb]

Hello @emcghee73

There is an option for this:

options("wru_data_wd" = TRUE)

If you download the files manually from the release and put them in the working directory, it won't access the internet. You'll need to download all the files manually:

[emcghee73]

Thanks! I tried this--downloading these four files and putting them in the working directory--and I got this error message (along with the code leading up to it):

setwd("Z:/projects/mcghee/Voter Files")
census20 <- readRDS("wru_census_2020.rds")
start.time <- proc.time()
options("wru_data_wd"=TRUE)
d2 <- predict_race(filter(d2, !is.na(surname), !is.na(tract), !is.na(county)),

•                census.geo="tract", census.data=census20, 
  

•                year="2020", age=F, sex=F, party="party", 
  

•                names.to.use="surname, first, middle", model="fBISG")
  

Using predict_race to obtain initial race prediction priors with BISG model
Proceeding with first, last, and middle name predictions...
There was an error retrieving dataCannot access release data for repo "kosukeimai/wru".
Proceeding with Census geographic data at tract level...
Using Census geographic data from provided census.data object...
State 1 of 1: CA
There was an error retrieving dataCannot access release data for repo "kosukeimai/wru".
Error in toupper(as.character(df$surname)) :
invalid input 'NUÑEZ' in 'utf8towcs'

proc.time() - start.time

That's farther than it made it before, but it's still giving the same basic error. Am I using the options command wrong?

[1beb]

Can you show me the str of your census object. Also, that error I think references a character encoding problem in your data. Can you also show us Sys.locale()?

[emcghee73]

I fixed the character encoding problem with stri_trans_general(LastName, "Latin-ASCII") and it the rest of it ran. So I now have imputed race probabilities and I think everything is ok? That said, I am still getting the error about accessing release data, and it seems like there are a lot of unmatched names (see below). Should I be worried about that?

Using predict_race to obtain initial race prediction priors with BISG model
Proceeding with first, last, and middle name predictions...
There was an error retrieving dataCannot access release data for repo "kosukeimai/wru".
Proceeding with Census geographic data at tract level...
Using Census geographic data from provided census.data object...
State 1 of 1: CA
There was an error retrieving dataCannot access release data for repo "kosukeimai/wru".
2363620 (11.4%) individuals' last names were not matched.
795341 (3.8%) individuals' first names were not matched.
797290 (3.8%) individuals' middle names were not matched.
There was an error retrieving dataCannot access release data for repo "kosukeimai/wru".
fBISG relies on MCMC; for reproducibility, I am setting RNG seed and returning it as attribute 'RNGseed'.
To silence this message, you can set a seed explicitly by defining the 'seed' element in the control list.
Forming Pr(race | location) tables from census data...

In case it's helpful, here's the output from str(census20). I couldn't get Sys.locale() to work--it said "could not find function Sys.locale()."

List of 1
$ CA:List of 7
..$ state : chr "CA"
..$ age : logi FALSE
..$ sex : logi FALSE
..$ year : chr "2020"
..$ block :'data.frame': 519723 obs. of 17 variables:
.. ..$ state : chr [1:519723] "CA" "CA" "CA" "CA" ...
.. ..$ county : chr [1:519723] "001" "001" "001" "001" ...
.. ..$ tract : chr [1:519723] "400100" "400100" "400100" "400100" ...
.. ..$ block : chr [1:519723] "1002" "1005" "1009" "1027" ...
.. ..$ P2_005N: num [1:519723] 0 0 0 49 144 0 65 17 55 68 ...
.. ..$ P2_006N: num [1:519723] 0 0 6 3 0 0 7 1 2 5 ...
.. ..$ P2_007N: num [1:519723] 0 0 0 0 0 0 0 0 0 0 ...
.. ..$ P2_008N: num [1:519723] 0 0 0 10 26 0 39 22 8 37 ...
.. ..$ P2_009N: num [1:519723] 0 0 0 0 0 0 0 0 0 0 ...
.. ..$ P2_010N: num [1:519723] 0 0 0 0 0 0 0 0 0 0 ...
.. ..$ P2_011N: num [1:519723] 0 0 2 2 16 0 9 7 6 3 ...
.. ..$ P2_002N: num [1:519723] 0 0 3 4 18 0 8 4 6 10 ...
.. ..$ r_whi : num [1:519723] 0.00 0.00 0.00 3.57e-06 1.05e-05 ...
.. ..$ r_bla : num [1:519723] 0.00 0.00 2.83e-06 1.42e-06 0.00 ...
.. ..$ r_his : num [1:519723] 0.00 0.00 1.93e-07 2.57e-07 1.16e-06 ...
.. ..$ r_asi : num [1:519723] 0.00 0.00 0.00 1.63e-06 4.25e-06 ...
.. ..$ r_oth : num [1:519723] 0.00 0.00 9.96e-07 9.96e-07 7.97e-06 ...
..$ tract :'data.frame': 9129 obs. of 16 variables:
.. ..$ state : chr [1:9129] "CA" "CA" "CA" "CA" ...
.. ..$ county : chr [1:9129] "001" "001" "001" "001" ...
.. ..$ tract : chr [1:9129] "400100" "400200" "400300" "400400" ...
.. ..$ P2_005N: num [1:9129] 1840 1383 3293 2676 1898 ...
.. ..$ P2_006N: num [1:9129] 140 39 544 283 617 ...
.. ..$ P2_007N: num [1:9129] 0 0 11 6 8 5 10 8 4 13 ...
.. ..$ P2_008N: num [1:9129] 538 190 540 386 304 125 406 574 243 566 ...
.. ..$ P2_009N: num [1:9129] 10 2 23 8 5 1 9 17 3 32 ...
.. ..$ P2_010N: num [1:9129] 49 14 37 29 38 42 52 40 24 49 ...
.. ..$ P2_011N: num [1:9129] 256 166 509 350 337 190 358 277 224 443 ...
.. ..$ P2_002N: num [1:9129] 205 207 547 374 437 188 633 500 374 977 ...
.. ..$ r_whi : num [1:9129] 0.000134 0.000101 0.00024 0.000195 0.000138 ...
.. ..$ r_bla : num [1:9129] 6.61e-05 1.84e-05 2.57e-04 1.34e-04 2.91e-04 ...
.. ..$ r_his : num [1:9129] 1.32e-05 1.33e-05 3.51e-05 2.40e-05 2.80e-05 ...
.. ..$ r_asi : num [1:9129] 8.96e-05 3.14e-05 9.20e-05 6.44e-05 5.05e-05 ...
.. ..$ r_oth : num [1:9129] 1.52e-04 8.97e-05 2.77e-04 1.92e-04 1.91e-04 ...
..$ county:'data.frame': 58 obs. of 15 variables:
.. ..$ state : chr [1:58] "CA" "CA" "CA" "CA" ...
.. ..$ county : chr [1:58] "001" "003" "007" "011" ...
.. ..$ P2_005N: num [1:58] 472277 801 139651 6941 455421 ...
.. ..$ P2_006N: num [1:58] 159499 10 3320 182 97994 ...
.. ..$ P2_007N: num [1:58] 4131 214 3050 280 2553 ...
.. ..$ P2_008N: num [1:58] 540511 12 10333 252 214520 ...
.. ..$ P2_009N: num [1:58] 13209 0 508 70 5720 ...
.. ..$ P2_010N: num [1:58] 10440 7 1184 92 8366 ...
.. ..$ P2_011N: num [1:58] 88537 76 13474 546 66453 ...
.. ..$ P2_002N: num [1:58] 393749 84 40112 13476 314900 ...
.. ..$ r_whi : num [1:58] 3.44e-02 5.84e-05 1.02e-02 5.06e-04 3.32e-02 ...
.. ..$ r_bla : num [1:58] 7.53e-02 4.72e-06 1.57e-03 8.59e-05 4.62e-02 ...
.. ..$ r_his : num [1:58] 2.53e-02 5.39e-06 2.57e-03 8.65e-04 2.02e-02 ...
.. ..$ r_asi : num [1:58] 9.05e-02 1.96e-06 1.77e-03 5.26e-05 3.60e-02 ...
.. ..$ r_oth : num [1:58] 0.051355 0.000148 0.00882 0.000457 0.038537 ...