Initial commit

.gitignore

@@ -0,0 +1 @@
+Preach_082212.csv

Makefile

@@ -0,0 +1,82 @@
+all: 			whatisEDA.md grdevices.md principles.md plotbase.md plottingsystems.md exploratorygraphs.md hclust.md example.md dimensionreduction.md colors.md ggplot2.md dplyr.md eda_checklist.md kmeans.md
+
+book_codefiles.zip:     all
+			./codefiles.R
+			cat codefiles_output.txt |xargs zip $@
+
+example.md: 			     example.Rmd
+				     knit.R $<
+				     perl -npi -e 's/```r/~~~~~~~~/' $@
+				     perl -npi -e 's/```/~~~~~~~~/' $@
+
+dimensionreduction.md: 		     dimensionreduction.Rmd
+				     knit.R $<
+				     perl -npi -e 's/```r/~~~~~~~~/' $@
+				     perl -npi -e 's/```/~~~~~~~~/' $@
+				     ## ./equation.pl $@
+				     R --no-save --args $@ < fixmath.R
+
+exploratorygraphs.md:	exploratorygraphs.Rmd
+			knit.R $<
+			perl -npi -e 's/```r/~~~~~~~~/' $@
+			perl -npi -e 's/```/~~~~~~~~/' $@
+
+hclust.md:		hclust.Rmd
+			knit.R $<
+			perl -npi -e 's/```r/~~~~~~~~/' $@
+			perl -npi -e 's/```/~~~~~~~~/' $@
+			##./equation.pl $@
+			R --no-save --args $@ < fixmath.R
+
+kmeans.md:		kmeans.Rmd
+			knit.R $<
+			perl -npi -e 's/```r/~~~~~~~~/' $@
+			perl -npi -e 's/```/~~~~~~~~/' $@
+
+dplyr.md:		dplyr.Rmd
+			knit.R $<
+			perl -npi -e 's/```r/~~~~~~~~/' $@
+			perl -npi -e 's/```/~~~~~~~~/' $@
+
+eda_checklist.md:	eda_checklist.Rmd
+			knit.R $<
+			perl -npi -e 's/```r/~~~~~~~~/' $@
+			perl -npi -e 's/```/~~~~~~~~/' $@
+
+ggplot2.md:		ggplot2.Rmd
+			knit.R $<
+			perl -npi -e 's/```r/~~~~~~~~/' $@
+			perl -npi -e 's/```/~~~~~~~~/' $@
+
+colors.md:		colors.Rmd
+			knit.R $<
+			perl -npi -e 's/```r/~~~~~~~~/' $@
+			perl -npi -e 's/```/~~~~~~~~/' $@
+
+cluster-example.md:	cluster-example.Rmd
+			knit.R $<
+			perl -npi -e 's/```r/~~~~~~~~/' $@
+			perl -npi -e 's/```/~~~~~~~~/' $@
+
+plotbase.md:		plotbase.Rmd
+			knit.R $<
+			perl -npi -e 's/```r/~~~~~~~~/' $@
+			perl -npi -e 's/```/~~~~~~~~/' $@
+
+principles.md:		principles.Rmd
+			knit.R $<
+			perl -npi -e 's/```r/~~~~~~~~/' $@
+			perl -npi -e 's/```/~~~~~~~~/' $@
+
+plottingsystems.md:	plottingsystems.Rmd
+			knit.R $<
+			perl -npi -e 's/```r/~~~~~~~~/' $@
+			perl -npi -e 's/```/~~~~~~~~/' $@
+
+grdevices.md:		grdevices.Rmd
+			knit.R $<
+			perl -npi -e 's/```r/~~~~~~~~/' $@
+			perl -npi -e 's/```/~~~~~~~~/' $@
+
+whatisEDA.md:		whatisEDA.Rmd
+			knit.R $<

_R_package_list.txt

@@ -0,0 +1,10 @@
+RColorBrewer
+dplyr
+impute
+readr
+ggplot2
+lubridate
+maps
+lattice
+datasets
+tsModel

cluster-example.Rmd

@@ -0,0 +1,181 @@
+# EDA Case Study - Understanding Human Activity with Smart Phones
+
+
+## Samsung Galaxy S3
+
+<img class=center src=images/samsung.png height=450>
+
+[http://www.samsung.com/global/galaxys3/](http://www.samsung.com/global/galaxys3/)
+
+
+
+
+## Samsung Data
+
+<img class=center src=images/ucisamsung.png height=450>
+
+[http://archive.ics.uci.edu/ml/datasets/Human+Activity+Recognition+Using+Smartphones](http://archive.ics.uci.edu/ml/datasets/Human+Activity+Recognition+Using+Smartphones)
+
+
+
+
+## Slightly processed data
+
+[Samsung data file]("https://dl.dropboxusercontent.com/u/7710864/courseraPublic/samsungData.rda")
+
+```{r loaddata,tidy=TRUE}
+load("data/samsungData.rda")
+names(samsungData)[1:12]
+table(samsungData$activity)
+```
+
+
+
+## Plotting average acceleration for first subject
+
+```{r processData,fig.height=4,fig.width=8,tidy=TRUE}
+par(mfrow=c(1, 2), mar = c(5, 4, 1, 1))
+samsungData <- transform(samsungData, activity = factor(activity))
+sub1 <- subset(samsungData, subject == 1)
+plot(sub1[, 1], col = sub1$activity, ylab = names(sub1)[1])
+plot(sub1[, 2], col = sub1$activity, ylab = names(sub1)[2])
+legend("bottomright",legend=unique(sub1$activity),col=unique(sub1$activity), pch = 1)
+```
+
+
+
+## Clustering based just on average acceleration
+
+<!-- ## source("http://dl.dropbox.com/u/7710864/courseraPublic/myplclust.R")  -->
+
+
+```{r dependson="processData",fig.height=5,fig.width=8}
+source("myplclust.R")
+distanceMatrix <- dist(sub1[,1:3])
+hclustering <- hclust(distanceMatrix)
+myplclust(hclustering, lab.col = unclass(sub1$activity))
+```
+
+
+
+
+## Plotting max acceleration for the first subject
+
+```{r ,dependson="processData",fig.height=5,fig.width=10}
+par(mfrow=c(1,2))
+plot(sub1[,10],pch=19,col=sub1$activity,ylab=names(sub1)[10])
+plot(sub1[,11],pch=19,col = sub1$activity,ylab=names(sub1)[11])
+```
+
+
+
+## Clustering based on maximum acceleration
+
+```{r dependson="processData",fig.height=5,fig.width=10}
+source("myplclust.R")
+distanceMatrix <- dist(sub1[,10:12])
+hclustering <- hclust(distanceMatrix)
+myplclust(hclustering,lab.col=unclass(sub1$activity))
+```
+
+
+
+
+
+## Singular Value Decomposition
+
+```{r svdChunk,dependson="processData",fig.height=5,fig.width=10,cache=TRUE,tidy=TRUE}
+svd1 = svd(scale(sub1[,-c(562,563)]))
+par(mfrow=c(1,2))
+plot(svd1$u[,1],col=sub1$activity,pch=19)
+plot(svd1$u[,2],col=sub1$activity,pch=19)
+```
+
+
+
+## Find maximum contributor
+
+```{r dependson="svdChunk",fig.height=5,fig.width=6,cache=TRUE,tidy=TRUE}
+plot(svd1$v[,2],pch=19)
+```
+
+
+
+
+##  New clustering with maximum contributer
+
+```{r dependson="svdChunk",fig.height=5,fig.width=8,cache=TRUE,tidy=TRUE}
+maxContrib <- which.max(svd1$v[,2])
+distanceMatrix <- dist(sub1[, c(10:12,maxContrib)])
+hclustering <- hclust(distanceMatrix)
+myplclust(hclustering,lab.col=unclass(sub1$activity))                             
+```
+
+
+
+
+##  New clustering with maximum contributer
+
+```{r dependson="svdChunk",fig.height=4.5,fig.width=4.5,cache=TRUE}
+names(samsungData)[maxContrib]                          
+```
+
+
+
+##  K-means clustering (nstart=1, first try)
+
+```{r kmeans1,dependson="processData",fig.height=4,fig.width=4}
+kClust <- kmeans(sub1[,-c(562,563)],centers=6)
+table(kClust$cluster,sub1$activity)
+```
+
+
+
+
+
+##  K-means clustering (nstart=1, second try)
+
+```{r dependson="kmeans1",fig.height=4,fig.width=4,cache=TRUE,tidy=TRUE}
+kClust <- kmeans(sub1[,-c(562,563)],centers=6,nstart=1)
+table(kClust$cluster,sub1$activity)
+```
+
+
+
+
+##  K-means clustering (nstart=100, first try)
+
+```{r dependson="kmeans1",fig.height=4,fig.width=4,cache=TRUE}
+kClust <- kmeans(sub1[,-c(562,563)],centers=6,nstart=100)
+table(kClust$cluster,sub1$activity)
+```
+
+
+
+
+
+##  K-means clustering (nstart=100, second try)
+
+```{r kmeans100,dependson="kmeans1",fig.height=4,fig.width=4,cache=TRUE,tidy=TRUE}
+kClust <- kmeans(sub1[,-c(562,563)],centers=6,nstart=100)
+table(kClust$cluster,sub1$activity)
+```
+
+
+
+##  Cluster 1 Variable Centers (Laying)
+
+```{r dependson="kmeans100",fig.height=4,fig.width=8,cache=FALSE,tidy=TRUE}
+plot(kClust$center[1,1:10],pch=19,ylab="Cluster Center",xlab="")
+```
+
+
+
+
+##  Cluster 2 Variable Centers (Walking)
+
+```{r dependson="kmeans100",fig.height=4,fig.width=8,cache=FALSE}
+plot(kClust$center[4,1:10],pch=19,ylab="Cluster Center",xlab="")
+```
+
+

codefiles.R

@@ -0,0 +1,20 @@
+#!/Users/rdpeng/bin/Rscript
+
+library(knitr)
+
+files <- readLines("Book.txt")
+files <- sub("md$", "Rmd", files, perl = TRUE)
+use <- file.exists(files) & files != "overview.Rmd"
+
+if(sum(use) == 0L) {
+        stop("no files")
+}
+
+files <- files[use]
+output <- character(length(files))
+
+for(i in seq_along(files)) {
+        output[i] <- knit(files[i], tangle = TRUE)
+}
+
+writeLines(output, "codefiles_output.txt")