[R] analyzing results from Tuesday's US elections
Bert Gunter
bgunter@4567 @end|ng |rom gm@||@com
Tue Nov 10 04:02:01 CET 2020
For those who are interested:
Very nice examples of (static) statistical graphics on election results can
be found here:
https://www.nytimes.com/interactive/2020/11/09/us/arizona-election-battleground-state-counties.html?action=click&module=Spotlight&pgtype=Homepage
Takes multidisciplinary teams and lots of hard work to produce, I would
guess.
Bert Gunter
"The trouble with having an open mind is that people keep coming along and
sticking things into it."
-- Opus (aka Berkeley Breathed in his "Bloom County" comic strip )
On Mon, Nov 9, 2020 at 4:46 PM Abby Spurdle <spurdle.a using gmail.com> wrote:
> RESENT
> INITIAL EMAIL, TOO BIG
> ATTACHMENTS REPLACED WITH LINKS
>
> I created a dataset, linked.
> Had to manually copy and paste from the NY Times website.
>
> > head (data, 3)
> STATE EQCOUNTY RMARGIN_2016 RMARGIN_2020 NVOTERS_2020
> SUB_STATEVAL_2016
> 1 Alabama Mobile 13.3 12 181783
> 0
> 2 Alabama Dallas -37.5 -38 17861
> 0
> 3 Alabama Tuscaloosa 19.3 15 89760
> 0
>
> > tail (data, 3)
> STATE EQCOUNTY RMARGIN_2016 RMARGIN_2020 NVOTERS_2020
> SUB_STATEVAL_2016
> 4248 Wyoming Uinta 58.5 63 9400
> 0
> 4249 Wyoming Sublette 63.0 62 4970
> 0
> 4250 Wyoming Johnson 64.3 61 4914
> 0
>
> > head (data [data [,1] == "Alaska",], 3)
> STATE EQCOUNTY RMARGIN_2016 RMARGIN_2020 NVOTERS_2020 SUB_STATEVAL_2016
> 68 Alaska ED 40 14.7 -24.0 82 1
> 69 Alaska ED 37 14.7 -1.7 173 1
> 70 Alaska ED 38 14.7 -0.4 249 1
>
> EQCounty, is the County or Equivalent.
> Several states, D.C., Alaska, Connecticut, Maine, Massachusetts, Rhode
> Island and Vermont are different.
> RMargin(s) are the republican percentages minus the democrate
> percentages, as 2 or 3 digit numbers between 0 and 100.
> The last column is 0s or 1s, with 1s for Alaska, Connecticut, Maine,
> Massachusetts, Rhode Island and Vermont, where I didn't have the 2016
> margins, so the 2016 margins have been replaced with state-levels
> values.
>
> Then I scaled the margins, based on the number of voters.
> i.e.
> wx2016 <- 1000 * x2016 * nv / max.nv
> (Where x2016 is equal to RMARGIN_2020, and nv is equal to NVOTERS_2020).
>
> There may be a much better way.
>
> And came up the following plots (linked) and output (follows):
>
> ---INPUT---
> PATH = "<PATH TO FILE>"
> data = read.csv (PATH, header=TRUE)
>
> #raw data
> x2016 <- as.numeric (data$RMARGIN_2016)
> x2020 <- as.numeric (data$RMARGIN_2020)
> nv <- as.numeric (data$NVOTERS_2020)
> subs <- as.logical (data$SUB_STATEVAL)
>
> #computed data
> max.nv <- max (nv)
> wx2016 <- 1000 * x2016 * nv / max.nv
> wx2020 <- 1000 * x2020 * nv / max.nv
> diffs <- wx2020 - wx2016
>
> OFFSET <- 500
> p0 <- par (mfrow = c (2, 2) )
>
> #plot 1
> plot (wx2016, wx2020,
> main="All Votes\n(By County, or Equivalent)",
> xlab="Scaled Republican Margin, 2016", ylab="Scaled Republican Margin,
> 2020")
> abline (h=0, v=0, lty=2)
>
> #plot 2
> OFFSET <- 200
> plot (wx2016, wx2020,
> xlim = c (-OFFSET, OFFSET), ylim = c (-OFFSET, OFFSET),
> main="All Votes\n(Zoomed In)",
> xlab="Scaled Republican Margin, 2016", ylab="Scaled Republican Margin,
> 2020")
> abline (h=0, v=0, lty=2)
>
> OFFSET <- 1000
>
> #plot 3
> J1 <- order (diffs, decreasing=TRUE)[1:400]
> plot (wx2016 [J1], wx2020 [J1],
> xlim = c (-OFFSET, OFFSET), ylim = c (-OFFSET, OFFSET),
> main="400 Biggest Shifts Towards Republican",
> xlab="Scaled Republican Margin, 2016", ylab="Scaled Republican Margin,
> 2020")
> abline (h=0, v=0, lty=2)
> abline (a=0, b=1, lty=2)
>
> #plot 4
> J2 <- order (diffs)[1:400]
> plot (wx2016 [J2], wx2020 [J2],
> xlim = c (-OFFSET, OFFSET), ylim = c (-OFFSET, OFFSET),
> main="400 Biggest Shifts Towards Democrat",
> xlab="Scaled Republican Margin, 2016", ylab="Scaled Republican Margin,
> 2020")
> abline (h=0, v=0, lty=2)
> abline (a=0, b=1, lty=2)
>
> par (p0)
>
> #most democrat
> I = order (wx2020)[1:30]
> cbind (data [I,], scaled.dem.vote = -1 * wx2020 [I])
>
> #biggest move toward democrat
> head (cbind (data [J2,], diffs = diffs [J2]), 30)
>
> ---OUTPUT---
> #most democrat
> > cbind (data [I,], scaled.dem.vote = -1 * wx2020 [I])
> STATE EQCOUNTY RMARGIN_2016 RMARGIN_2020
> NVOTERS_2020 SUB_STATEVAL_2016 scaled.dem.vote
> 229 California Los Angeles -49.3 -44
> 3674850 0 44000.000
> 769 Illinois Cook -53.1 -47
> 1897721 0 24271.164
> 4073 Washington King -48.8 -53
> 1188152 0 17135.953
> 3092 Pennsylvania Philadelphia -67.0 -63
> 701647 0 12028.725
> 215 California Alameda -63.5 -64
> 625710 0 10897.163
> 227 California Santa Clara -52.1 -49
> 726186 0 9682.875
> 238 California San Diego -19.7 -23
> 1546144 0 9676.942
> 2683 New York Brooklyn -62.0 -49
> 693937 0 9252.871
> 2162 Minnesota Hennepin -34.9 -43
> 753716 0 8819.350
> 2074 Michigan Wayne -37.1 -37
> 863382 0 8692.908
> 2673 New York Manhattan -76.9 -70
> 446861 0 8511.986
> 221 California San Francisco -75.2 -73
> 413642 0 8216.898
> 3495 Texas Dallas -26.1 -32
> 920772 0 8017.934
> 1741 Maryland Prince George's -79.7 -80
> 365857 0 7964.559
> 510 Florida Broward -34.9 -30
> 959418 0 7832.303
> 3057 Oregon Multnomah -56.3 -61
> 458395 0 7609.044
> 3563 Texas Travis -38.6 -45
> 605034 0 7408.882
> 565 Georgia DeKalb -62.9 -67
> 369341 0 6733.839
> 3942 Virginia Fairfax -35.8 -42
> 578931 0 6616.624
> 492 D.C. D.C. -86.4 -87
> 279152 0 6608.766
> 562 Georgia Fulton -40.9 -46
> 522050 0 6534.770
> 230 California Contra Costa -43.0 -48
> 498340 0 6509.196
> 2674 New York Queens -53.6 -39
> 597928 0 6345.617
> 257 Colorado Denver -54.8 -64
> 350606 0 6106.041
> 2677 New York Bronx -79.1 -66
> 329638 0 5920.271
> 3530 Texas Harris -12.3 -13
> 1633671 0 5779.208
> 1718 Maryland Montgomery -55.4 -57
> 369405 0 5729.781
> 2888 Ohio Cuyahoga -35.2 -34
> 605268 0 5599.987
> 2745 North Carolina Mecklenburg -29.4 -35
> 565980 0 5390.506
> 2894 Ohio Franklin -25.8 -31
> 606022 0 5112.231
>
> #biggest move toward democrat
> > head (cbind (data [J2,], diffs = diffs [J2]), 30)
> STATE EQCOUNTY RMARGIN_2016 RMARGIN_2020
> NVOTERS_2020 SUB_STATEVAL_2016 diffs
> 1751 Massachusetts Boston -26.8 -67.00
> 273133 1 -2987.8625
> 113 Arizona Maricopa 2.8 -2.00
> 2046295 0 -2672.8209
> 3531 Texas Tarrant 8.6 -0.16
> 830104 0 -1978.7776
> 2162 Minnesota Hennepin -34.9 -43.00
> 753716 0 -1661.3194
> 3564 Texas Collin 16.7 5.00
> 486917 0 -1550.2480
> 3495 Texas Dallas -26.1 -32.00
> 920772 0 -1478.3065
> 238 California San Diego -19.7 -23.00
> 1546144 0 -1388.4309
> 563 Georgia Gwinnett -5.8 -18.00
> 413166 0 -1371.6547
> 3565 Texas Denton 20.0 8.00
> 416610 0 -1360.4147
> 4073 Washington King -48.8 -53.00
> 1188152 0 -1357.9434
> 564 Georgia Cobb -2.2 -14.00
> 393340 0 -1263.0208
> 2075 Michigan Oakland -8.1 -14.00
> 778418 0 -1249.7561
> 291 Colorado Jefferson -6.9 -19.00
> 376430 0 -1239.4528
> 292 Colorado El Paso 22.3 11.00
> 375058 0 -1153.2866
> 2321 Missouri St. Louis County -16.2 -24.00
> 528107 0 -1120.9259
> 3563 Texas Travis -38.6 -45.00
> 605034 0 -1053.7077
> 277 Colorado Arapahoe -14.1 -25.00
> 346740 0 -1028.4681
> 2744 North Carolina Wake -20.2 -26.00
> 624049 0 -984.9339
> 3942 Virginia Fairfax -35.8 -42.00
> 578931 0 -976.7398
> 1116 Kansas Johnson 2.6 -8.00
> 338343 0 -975.9407
> 3562 Texas Bexar -13.4 -18.00
> 757667 0 -948.4110
> 2077 Michigan Kent 3.1 -6.00
> 359915 0 -891.2545
> 257 Colorado Denver -54.8 -64.00
> 350606 0 -877.7434
> 110 Arizona Pima -13.6 -20.00
> 501058 0 -872.6264
> 2625 New Jersey Monmouth 9.3 -1.60
> 292654 0 -868.0432
> 2745 North Carolina Mecklenburg -29.4 -35.00
> 565980 0 -862.4809
> 3567 Texas Williamson 9.7 -1.30
> 287696 0 -861.1660
> 2894 Ohio Franklin -25.8 -31.00
> 606022 0 -857.5355
> 203 California Riverside -5.4 -11.00
> 558759 0 -851.4770
> 3966 Virginia Virginia Beach 3.5 -8.00
> 253477 0 -793.2257
>
> DISCLAIMER:
> I can not guarantee the accuracy of this data, or any conclusions.
>
> NOTE:
> Reiterating, several states used state-level values for 2016.
> (So, the Boston value above, may be off).
>
> Monospaced fonts are required for reading the contents of this email.
>
> LINKS:
>
> https://sites.google.com/site/spurdlea/temp_election
>
> https://sites.google.com/site/spurdlea/exts/election_data.txt
>
[[alternative HTML version deleted]]
More information about the R-help
mailing list