How to choose only top five options in scatterplot

Question

I want to choose only the top five options in a scatterplot. Here's the code:

library(dslabs)

data(polls_us_election_2016)

polls <- polls_us_election_2016 %>%
  filter(state == "U.S." & enddate >= "2016-10-31" &
           (grade %in% c("A+","A","A-","B+") | is.na(grade)))

polls <- polls %>%
  mutate(spread = rawpoll_clinton/100 - rawpoll_trump/100)

d_hat <- polls %>%
  summarize(d_hat = sum(spread * samplesize) / sum(samplesize)) %>%
  pull(d_hat)

p_hat <- (d_hat+1)/2

moe <- 1.96 * 2 * sqrt(p_hat * (1 - p_hat) / sum(polls$samplesize))

polls %>% ggplot(aes(spread, pollster)) +
  geom_point(color="black", binwidth = .01)

This is the output graph:

The following is the desired graph:

score 0 · Answer 1 · answered Jun 21 '22 at 11:28

The pipe nest %>% head(5) %>% unnest gives you the first 5 groups of rows in the order they occur in the data frame. Then, ggplot2 will sort them in an alpha lexical way:

library(dslabs)
library(tidyverse)

data(polls_us_election_2016)

polls <- polls_us_election_2016 %>%
  filter(state == "U.S." & enddate >= "2016-10-31" &
           (grade %in% c("A+","A","A-","B+") | is.na(grade)))

polls <- polls %>%
  mutate(spread = rawpoll_clinton/100 - rawpoll_trump/100)

d_hat <- polls %>%
  summarize(d_hat = sum(spread * samplesize) / sum(samplesize)) %>%
  pull(d_hat)

p_hat <- (d_hat+1)/2

moe <- 1.96 * 2 * sqrt(p_hat * (1 - p_hat) / sum(polls$samplesize))

polls %>%
  nest(-pollster) %>%
  head(5) %>%
  unnest(data) %>%
  ggplot(aes(spread, pollster)) +
  geom_point(color="black", binwidth = .01) 
#> Warning: All elements of `...` must be named.
#> Did you want `data = -pollster`?
#> Warning: Ignoring unknown parameters: binwidth

^{Created on 2022-06-21 by the reprex package (v2.0.0)}

this does not correspond to the desired output. Also, I feel the way via nest/unnest seems really very involved. — tjebo, Jun 21 '22 at 12:37
Please specify in the question on how to define the 'top five' — danlooo, Jun 21 '22 at 14:09

score 0 · Answer 2 · answered Jun 21 '22 at 13:40

This is essentially a "how to chose groups with n most rows/ values" problem.

Here one approach by first summarising and using this information to filter the orginal data.

library(dslabs)
library(tidyverse)

data(polls_us_election_2016)

polls <- polls_us_election_2016 %>%
  filter(state == "U.S." & enddate >= "2016-10-31" &
           (grade %in% c("A+","A","A-","B+") | is.na(grade)))

polls <- polls %>%
  mutate(spread = rawpoll_clinton/100 - rawpoll_trump/100)

polls %>% 
##count number within each group and slice the first 5
count(pollster) %>%
slice_max(n, n = 5) %>%
## merge back with your original data
semi_join(polls, ., by = "pollster") %>%
## plot as suggested by yourself
ggplot(aes(spread, pollster)) +
  geom_point(color="black", binwidth = .01)

How to choose only top five options in scatterplot

2 Answers2