Chapter 4 Tags and Percentage
Data Visualization and EDA are the top most Tag contributors for the hidden gems. Jonathan Bouchet and Ramshankar Yadhunath have data visualization and eda as their most popular tags. Vopani’s popular tags include beginner , tpu , tensorflow. Bojan Tunguz’s most popular tags include gpu , xgboost
TotalNoOfRows = nrow(kernels_gems_tags)
p1 <- kernels_gems_tags %>%
group_by(Slug) %>%
summarise(Percentage = n()/TotalNoOfRows *100) %>%
arrange(desc(Percentage)) %>%
head(10) %>%
ungroup() %>%
mutate(Slug = reorder(Slug,Percentage))
p1 %>%
filter(!is.na(Slug)) %>%
mutate(foo = Percentage > 9 ) %>%
ggplot(aes(x = Slug,y = Percentage, fill = foo )) +
geom_bar(stat='identity',colour="white") +
scale_fill_manual(values = c(fillColor2,fillColor)) +
geom_text(aes(x = Slug, y = 1,
label = paste0("( ",round(Percentage,2)," %)",sep="")),
hjust=0, vjust=.5, size = 6, colour = 'black',
fontface = 'bold') +
labs(x = 'Slug',
y = 'Percentage',
title = 'Tags and Percentage') +
guides(fill=guide_legend(title="Tags Percentage")) +
coord_flip() +
theme_fivethirtyeight(base_size = 15) +
theme(legend.position = "none")
4.1 Jonathan Bouchet Tags
getTagsAuthor <- function(author) {
kernels_gems_tags_jb = kernels_gems_tags %>%
filter(author_name == author)
TotalNoOfRows = nrow(kernels_gems_tags_jb)
p1 <- kernels_gems_tags_jb %>%
group_by(Slug) %>%
summarise(Percentage = n()/TotalNoOfRows *100) %>%
arrange(desc(Percentage)) %>%
head(10) %>%
ungroup() %>%
mutate(Slug = reorder(Slug,Percentage))
p1 %>%
filter(!is.na(Slug)) %>%
mutate( foo = Percentage > 10 ) %>%
ggplot(aes(x = Slug,y = Percentage, fill = foo )) +
geom_bar(stat='identity',colour="white") +
scale_fill_manual(values = c(fillColor2,fillColor)) +
geom_label(aes(label = paste0("( ",round(Percentage,2)," %)",sep=""))) +
labs(x = 'Slug',
y = 'Percentage',
title = 'Tags and Percentage') +
guides(fill=guide_legend(title="Tags Percentage")) +
coord_flip() +
theme_fivethirtyeight(base_size = 15) +
theme(legend.position = "none")
}
getTagsAuthor("Jonathan Bouchet")