Find the cumulative number of albums an artist has released and predict future counts.

album_counts.R

# https://en.wikipedia.org/wiki/The_Mountain_Goats_discography
# Scraping
library("rvest")
library("ggplot2")
library("ggthemes")
library("stringr")
library("tidyverse")

tbls <- "https://en.wikipedia.org/wiki/The_Mountain_Goats_discography" %>%
  read_html %>% html_nodes("table")
# Shows all tables
tbls

# Matched more than one, so using the first result when assigning value to df
tbls[grep("Zopilote Machine",tbls,ignore.case = T)]

# Get those albums
df <- html_table(tbls[grep("Zopilote Machine",tbls,ignore.case = T)[1]],fill = T)[[1]]

# Find a sequence of four numbers for year
df$Year <- str_extract_all(df$`Album details`,"[0-9]{4}") %>% as.numeric

# Drop first row because it is a duplicate of the header (wiki table thing)
df <- df[-1,c("Title","Album details","Year")]
df_plot <- min(df$Year):max(df$Year) %>% lapply(function(year){
  sum(df$Year<=year)
}) %>% 
  unlist %>% 
  data.frame(Year = min(df$Year):max(df$Year)) %>% select(Year = "Year",Albums=".")

ggplot(df_plot,aes(x=Year,y=Albums)) + 
  geom_bar(stat="identity") + theme_fivethirtyeight() +
  scale_x_continuous(breaks=seq(min(df_plot$Year),max(df_plot$Year)+40,5)) +
  xlim(min(df_plot$Year)-1,2040) +
  stat_smooth(method="lm",fullrange=TRUE) + 
  # Give back those axis that theme_fivethirtyeight removed
  theme(axis.title = element_text()) +
  labs(title = "Mountain Goats Albums by Year",
       subtitle = "Studio Album Releases from https://en.wikipedia.org/wiki/The_Mountain_Goats_discography",
       caption = "\n@appupio") +
  ylab("Number of Albums\n") + xlab("Year")