text analysis for breast pump user data

kanarinka · Dec 15, 2014 · ec69e7b · ec69e7b
1 parent 9c9edd6
commit ec69e7b
Show file tree

Hide file tree

Showing 8 changed files with 513 additions and 0 deletions.
diff --git a/.gitignore b/.gitignore
@@ -0,0 +1,5 @@
+.Rproj.user
+.Rhistory
+.RData
+
+*.Rproj
diff --git a/textanalysis/bigrams.csv b/textanalysis/bigrams.csv
diff --git a/textanalysis/frequencyByCategory-IWouldLove.png b/textanalysis/frequencyByCategory-IWouldLove.png
diff --git a/textanalysis/frequencyByCategory.png b/textanalysis/frequencyByCategory.png
diff --git a/textanalysis/grouptrigrams.R b/textanalysis/grouptrigrams.R
@@ -0,0 +1,20 @@
+library(plyr)
+library(ggplot2)
+df = read.csv('trigrams_categories.csv')
+#filter those without categories
+df=df[df$category!="",] 
+
+#sum categories
+df = ddply(df, 'category', function(x) c(freq=sum(x$freq), separate_trigrams=nrow(x)))
+df=df[df$category!="",] 
+
+# Possibly remove "IWOULDLOVE"
+#df=df[df$category!="IWOULDLOVE",] 
+
+#order by frequency
+df[ order(df$freq), ]
+
+#plot to verify we did it right
+p <- ggplot(df, aes(y=freq,fill=category))
+df$category <- reorder(df$category, -df$freq)
+p + ylab("Frequency of Mention")+xlab("")+geom_bar(aes(x=category), data=df, stat="identity") +theme(axis.text=element_text(angle=0)) + coord_flip()