BD03

setwd("C:/Users/subas/Syncplicity/MyProjects_IMP/MY_Papers_V2/TRB 2020/00000000 FINALz/0050 GREG GRIFFIN- autonomous survey")
book_words <- read.csv("Book_words2.csv")
library(dplyr)

## Warning: package 'dplyr' was built under R version 3.6.2

## 
## Attaching package: 'dplyr'

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

library(tidyr)

## Warning: package 'tidyr' was built under R version 3.6.2

library(tidytext)

## Warning: package 'tidytext' was built under R version 3.6.2

library(ggplot2)

## Warning: package 'ggplot2' was built under R version 3.6.2

mm= book_words %>% 
  arrange(desc(tf_idf)) %>% 
  group_by(Ques) %>% slice(1:15)
dim(mm)

## [1] 105   7

head(mm)

## # A tibble: 6 x 7
## # Groups:   Ques [1]
##    Ques word            n total     tf   idf tf_idf
##   <int> <fct>       <int> <int>  <dbl> <dbl>  <dbl>
## 1     1 redacted        4   139 0.0288  1.25 0.0361
## 2     1 years           4   139 0.0288  1.25 0.0361
## 3     1 done            2   139 0.0144  1.95 0.0280
## 4     1 responsible     2   139 0.0144  1.95 0.0280
## 5     1 ago             3   139 0.0216  1.25 0.0270
## 6     1 started         3   139 0.0216  1.25 0.0270

ggplot(mm, aes(tf_idf, word)) +
  geom_segment(aes(x = 0, y = reorder(word, tf_idf), xend = tf_idf, yend = word), color = "grey50") +
  geom_point() +
  facet_wrap(~ Ques, scales = "free_y")+theme_bw(base_size = 18)