This repository created for an assignment as part of the Coursera course Getting and Cleaning data.

The data set present is a tidier data that is downloaded from https://d396qusza40orc.cloudfront.net/getdata%2Fprojectfiles%2FUCI%20HAR%20Dataset.zip.

• Dataset: activity_subject_summary.txt

• Description: This is the data set with the average of each measurement for each activity and each subject.

• Code Used for Transformation: run_analysis.R

The analysis function reads the files from the actual dataset and transforms to our needs to make tidier data set which can be used for analysis.

• Read the features list from features.txt file. There are some duplicate columns which will cause problems while filtering based on column name. So a unique columnn concatenating both the column values is created. This unique column will be used as header for the dataset.

features <- tbl_df(read.table("UCI HAR Dataset/features.txt", stringsAsFactors=FALSE, col.names = c("fid", "fnm"))) %>% mutate(varname = paste(fid,fnm))

• Load activities table with the activity identifier and its description. This will be used to resolve the id column from the actual data to its description.

• Data provided is scattered in three different files. These files are identified with format x_* , subject*, and y_*. Use feature list table created above to name the columns. This is done in two steps once for test data and then for train data.

• Get the test data from X_test, subject_test, y_test and column bind them all to get single table with test data.

xtest <- read.table("UCI HAR Dataset/test/X_test.txt", stringsAsFactors=FALSE)

names(xtest) <- features$varname

xtestsubject <- read.table("UCI HAR Dataset/test/subject_test.txt", stringsAsFactors=FALSE, col.names = c("subject"))

xtestactivity <- read.table("UCI HAR Dataset/test/y_test.txt", stringsAsFactors=FALSE, col.names = c("activity"))

xtest <- cbind(xtestsubject, xtestactivity, xtest)

• Get the test data from X_train, subject_train, y_train and column bind them all to get single table with train data.

xtrain <- read.table("UCI HAR Dataset/train/X_train.txt", quote=""", comment.char="", stringsAsFactors=FALSE)

names(xtrain) <- features$varname

xtrainsubject <- read.table("UCI HAR Dataset/train/subject_train.txt", quote=""", comment.char="", stringsAsFactors=FALSE, col.names = c("subject"))

xtrainactivity <- read.table("UCI HAR Dataset/train/y_train.txt", quote=""", comment.char="", stringsAsFactors=FALSE, col.names = c("activity"))

xtrain <- cbind(xtrainsubject, xtrainactivity, xtrain)

• Combine the test and train data using rowbind and select only those columsn that end with std() and mean() to get the standard deviation and mean of individual measurements.

xall <- tbl_df(rbind(xtest, xtrain)) %>% select (1, 2, ends_with("-std()"), ends_with("-mean()") )

• Join the result data with activities table to resolve the activityId to actual name and remove the activity column as it is redundant. Make the names removing unnecessary punctuations and meaning less numbers. After this point of time the data frame x3 is the tidier data that can be used for any further analysis.

x3 <- select(tbl_df(merge(activities, xs)), -activity)

names(x3) <- gsub("\(\)", "",gsub("^[0-9]{3} ", "",names(x3)))

• As there are multiple obervations for particular activity and subject. To get mean over each combination we set group_by over x3 on activityname and subject. Then summarise the data with mean function group by activityname and subject. Write the end result to a file.

activity_subject_summary <- group_by(x3, activityname, subject) %>% summarise_each(c("mean"))

write.table(activity_subject_summary, file = "activity_subject_summary.txt", row.names=FALSE)