In COMHIS/estc: English Short Title Catalogue (ESTC) Metadata Toolkit

# Scotland publishing summaries
# Read the preprocessed ESTC data table and load tools:
# Load libraries
library(ggplot2)
library(tidyr)
library(reshape2)
library(estc)
library(bibliographica)
library(dplyr)

# Complete data
dfo <- df.preprocessed
dfo$unity <- rep(1, nrow(dfo))

# Pick Scotland documents only
selected_place <- "Edinburgh"
sel.country <- "Scotland"
df <- filter(dfo, country == sel.country & publication_year >= 1470)

publications <- tapply(df$unity, list(df$publication_decade, df$publication_place), sum)
publications[is.na(publications)] <- 0 # Set NAs to 0
publications <- publications/10 # Instead of decadal sum, use average annual output 
dfm <- melt(publications) 
names(dfm) <- c("Time", "Place", "Documents")
dfm <- filter(dfm, Place == selected_place)

df <- transform(dfm, date = as.character(Time))
varname <- selected_place
dfs <- spread(df, Place, Documents)
dfs$date <- as.numeric(as.character(dfs$date))
dfs$varname <- dfs[[varname]]
p <- ggplot(dfs, aes(x = date, y = varname))
p <- p + geom_bar(stat = "identity", fill = "darkgray", col = "black")
p <- p + scale_x_continuous(breaks=seq(min(dfm$Time), max(dfm$Time), 20))
p <- p + ggtitle(paste("Documents published in ", varname))
p <- p + ylab("Documents / Year")
print(p)

# ------------------------------------------------------------

# Use 5 year intervals
df <- filter(dfo, country == sel.country & publication_year >= 1470)
timeinterval <- 5
df$publication.timeunit <- round(df$publication_year/timeinterval)*timeinterval 
publications <- tapply(df$unity, list(df$publication.timeunit, df$publication_place), sum)
publications[is.na(publications)] <- 0 # Set NAs to 0
publications <- publications/timeinterval # Instead of decadal sum, use average annual output 
dfm <- melt(publications) 
names(dfm) <- c("Time", "Place", "Documents")
dfm <- filter(dfm, Place == selected_place)
dfm <- transform(dfm, date = as.character(Time))
dfs <- spread(dfm, Place, Documents)
dfs$date <- as.numeric(as.character(dfs$date))
dfs$varname <- dfs[[selected_place]]

# Custom highlight for specific time intervals
rect_left <- c(min(na.omit(dfs$date)),
               1642, 1651+1, 
               1660, 1660+1,
               1688, 1689+1,
               1706, 1707+1,
               1776, 1776+1,
               max(na.omit(dfs$date)))
  rectangles <- data.frame(
    xmin = rect_left[-length(rect_left)],
    xmax = rect_left[-1],
    ymin = min(dfs$varname),
    ymax = max(dfs$varname))
  rectangles$shade <- rep(c("White", "Highlight"), length = nrow(rectangles))


# Draw Figure
theme_set(theme_bw(20))
p <- ggplot()
p <- p + geom_rect(data = rectangles, 
       aes(xmin=xmin, xmax=xmax, ymin=ymin, ymax=ymax, fill=shade), alpha=0.8) + 
         scale_fill_manual(values = c("gray", "white")) +
         guides(fill = "none") 
p <- p + geom_line(data = dfs, aes(x = date, y = varname), col = "black")
p <- p + geom_point(data = dfs, aes(x = date, y = varname), col = "black")
p <- p + scale_x_continuous(breaks=seq(min(dfs$date), max(dfs$date), 20))
p <- p + ggtitle(paste("Publishing activity in ", selected_place))
p <- p + ylab("Documents / Year")
print(p)


# -----------------------------------------------------------------------

highlight_years <- c(1642:1651, 1660, 1688:1689, 1706:1707, 1776)
df$highlight <- df$publication_year %in% highlight_years
hits <- c()
for (i in 1:nrow(dfs)) {
  hits[[i]] <- any((dfs$date[[i]] - round(timeinterval/2)):(dfs$date[[i]] + round(timeinterval/2)) %in% highlight_years)
}
dfs$highlight <- hits

theme_set(theme_bw(20))
p <- ggplot()
p <- p + geom_bar(data = dfs, aes(x = date, y = varname, fill = highlight), stat = "identity", alpha = 0.5, col = "black")
p <- p + scale_x_continuous(breaks=seq(min(dfs$date), max(dfs$date), 20))
p <- p + ggtitle(paste("Publishing activity in ", selected_place))
p <- p + ylab("Documents / Year")
p <- p + guides(fill = "none") 
p <- p + scale_fill_manual(values = c("gray", "black"))
print(p)

# use average annual output therefore divide by 10
dfm <- df %>% group_by(publication_decade, publication_place) %>%
              summarize(paper = sum(paper, na.rm = T)/10, n = n()/10)

theme_set(theme_bw(20))
p <- ggplot(dfm, aes(x = publication_decade, y = n, color = publication_place))
p <- p + geom_point()
p <- p + geom_line()
p <- p + xlab("Decade") + ylab("Publications per year (n)")
p <- p + ggtitle("Published documents")
p <- p + guides(color = guide_legend(title="Publication place"))
print(p)

theme_set(theme_bw(20))
p <- ggplot(dfm, aes(x = publication_decade, y = paper, color = publication_place))
p <- p + geom_point()
p <- p + geom_line()
p <- p + xlab("Decade") + ylab("Annual paper consumption")
p <- p + ggtitle("Paper consumption")
p <- p + guides(color = guide_legend(title="Publication place"))
print(p)