ProStudenty.R

# https://developers.facebook.com/tools/explorer/

install.packages("vioplot")
install.packages("plyr")

library(plyr)
library(vioplot)
library(Rfacebook)
library(xlsx)

token <- "CAACEdEose0cBAIdf3CeSo9jb5R6ksiZC4Ez4CtyBV1JgkRRHjEGZCt9gc1rKl7tqYsUOyMOXvaXAJqZBU86GWGSkBWomre0NnvAMAKdkCLU1nZBb3wEV6ZBHZBvDddgOUVXVTtmIqGlrrinY1F2MXZAOOApIfyDQ31WNz4R4UMnQeP9cCHT2ZAvQZCeypbsEsu0cMDvwX00ugfbF1HQFOwdXpZCljmdc3CpZC0ZD"
x = 300
name <- "hnutiusvit.cz"
page <- getPage(name, token, n = x)
write.csv(page, "hnutiusvit.cz.csv")
filename <- paste0(name, ".xls", sep="")
write.xlsx(page, filename, sheetName=name)


# druhy
page$likes_count_cat <- cut(page$likes_count,
breaks=c(-Inf, quantile(page$likes_count, .20), quantile(page$likes_count, .80), Inf), labels=c("low", "normal", "high"))

page$comments_count_cat <- cut(page$comments_count,
breaks=c(-Inf, quantile(page$comments_count, .20), quantile(page$comments_count, .80), Inf),
labels=c("low", "normal", "high"))

cases <- data.frame(page$type, page$likes_count_cat, page$comments_count_cat)

pravidla <- apriori(cases, 
                     parameter=list(support=0.02, confidence=0.05),
                     appearance = list(rhs=c("page.likes_count_cat=high"),
                                       default="lhs"),
                     control = list(verbose=F))

pravidla.sorted <- sort(pravidla, by="lift")
inspect(pravidla.sorted)

subset.matrix <- is.subset(pravidla.sorted, pravidla.sorted)
subset.matrix[lower.tri(subset.matrix, diag=T)] <- NA
redundant <- colSums(subset.matrix, na.rm=T) >= 1
# remove redundant rules
rules.pruned <- pravidla.sorted[!redundant]
inspect(rules.pruned)



# zjsiteni typu
class(page)

# vypis ukazky
head(page$type)

# zakladni stats
summary(page$likes_count)

# histogram
 hist(page$likes_count)

# boxplot
boxplot(page$likes_count)

boxplot(page$likes_count, outline=FALSE)
boxplot(page$likes_count, page$comments_count, outline=FALSE)

boxplot(page$likes_count, page$comments_count, page$shares_count, names=c("likes","comments","shares"), outline=FALSE)

table(page$type)

pie(druhy)

vioplot(page$likes_count, page$comments_count, page$shares_count, names=c("likes","comments","shares"))

plot(page$likes_count, page$comments_count)

plot(page$likes_count, page$comments_count, xlab="pocet likes", ylab="pocet komentaru", main="HNUTI")

postPage <- subset(page, type=="link")

page$type[page$likes_count > 200]

quantile(page$likes_count,.95)

table(page$type[page$likes_count>85])

page$time <-  format(as.POSIXlt(page$created_time))
likes_by_day <-  aggregate(likes_count ~ time, page, sum)
colnames(likes_by_day) <- c("date", "sum")
likes_by_day$date <- as.Date(likes_by_day$date)

page$months <- as.numeric(format(as.POSIXlt(page$time), format = "%y%m"))
likes_by_date_months <-  aggregate(likes_count ~  months + time, page, sum)

boxplot(likes_by_date_months$likes_count ~ likes_by_date_months$months, outline=FALSE, main="Dynamika like na Hnuti Usvit")
plot(as.Date(likes_by_day$time),likes_by_day$likes_count)
plot(as.Date(likes_by_day$time),likes_by_day$likes_count, type="h")
plot(as.Date(likes_by_day$time),likes_by_day$likes_count, type="h", xlab="datum", ylab="pocet likes", main="Sramkova Sobotka")

par(mfrow=c(1,2))
 plot(as.Date(likes_by_day$time),likes_by_day$likes_count, type="h", xlab="datum", ylab="pocet likes", main="Sramkova Sobotka")
 boxplot(likes_by_date_months$likes_count ~ likes_by_date_months$months, outline=FALSE, main="Dynamika like na Sramkova Sobotka")


png(filename="obrazek.png", width=800, height=600)
 boxplot(likes_by_date_months$likes_count ~ likes_by_date_months$months, outline=FALSE, main="Dynamika like na Sramkova Sobotka")
dev.off()

datum <- data.frame(seq(from = as.Date("2014-09-30"), to = as.Date("2015-03-01"), by = 1))
colnames(datum) <- c("date")

sjednocena <- join(datum, likes_by_day, type="left")
barplot(sjednocena$sum)

# http://thinktostart.com/cluster-twitter-data-with-r-and-k-means/