Getting And Cleaning Data course project

Files:

• README.md: Readme file. Describe how script works.
• run_analysis.R: Script for this course project
• tidy.csv.txt: The result tidy file. It's a csv file, so you can just remove the '.txt' then view it in excel.
• Codebook.md: Codebook for tidy dataset.

run_analysis.R

Usage:

1. Download data file from: https://d396qusza40orc.cloudfront.net/getdata%2Fprojectfiles%2FUCI%20HAR%20Dataset.zip
2. Unzip this file to your working dir. So you got the 'UCI HAR Dataset' folder in your working directory.
3. Use 'source("run_analysis.R")' to run the script. It will take you about one minute.
4. You got the tidy dataset 'tidy.csv.txt'.

How script working:

This script will archive these goals:

• Goal 1. Merges the training and the test sets to create one data set.
• Goal 2. Extracts only the measurements on the mean and standard deviation for each measurement.
• Goal 3. Uses descriptive activity names to name the activities in the data set.
• Goal 4. Appropriately labels the data set with descriptive variable names.
• Goal 5. Creates a second, independent tidy data set with the average of each variable for each activity and each subject.

I was trying to keep this script simple and short. So it's working in order 1, 4, 3, 2, 5. It will save some lines of code in this order(one goal one sentence).

First, I create some virables for filenames, it will makes your code looks beautiful.

And if you unzip your data to somewhere else, you juse need to change data.dir to point to your data dir. Easy to change. :)

Goal 1: Merges the training and the test sets to create one data set.

I use read.table here, because I don't need to change any default option here. See the document for read.table, you will see head = FALSE, sep = "", it's perfect for this project, it's the shortest solution. And read.table will return a data.frame, so I can use cbind to bind subject and activity together directly. Why? Because cbind(read.table("filename"), read.table("filename2")) means cbind(data.frame returned by read.table("filename"), and data.frame returned by read.table("filename2")). And then the same for rbind, it will connect train data and test data together. Because rbind(cbind(...), cbind(...)) means rbind(data.frame returnd by the 1st cbind, and data.frame returned by the 2nd cbind). Finally I got a big and complete data.frame. The code look nice right?

Goal 4: Appropriately labels the data set with descriptive variable names.

After Goal 1 finished, all datas are ready now. It's a good time to add names to columns. Why here? Because I can use column name for reference later if I done it here. It means I can use data$activity instead of data$Vxxx or data[[xxx]]. I read the features file and add 'subject' and 'activity' as colnames. Now every column have their name. colnames(data) will return or set the colnames, I will set the colnames. read.table will return a data frame, so read.table(...)$V2 will return the 2nd col of that data frame. If you see the file, you will know it's a string vector(with option stringsAsFactors = FALSE in read.table). Great. This sentence means I'll read the file 'features.txt' as colnames for data. But data have two more column I added from Goal 1, so I need to add these column names to that vector returned by read.table()$V2. To combine vectors, use c().

Goal 3: Uses descriptive activity names to name the activities in the data set.

Activity file is arranged in right order, so I can use them directly. Use data$activity as index, replace data$activity column with descriptive names. If you open subject_train.txt or subject_test.txt file, you will see they're all numbers. Those numbers and the line numbers of activity file are concordant. That's a good news, because I can forget about the merge process, and use data$activity as index directly to index the descriptive activity names inside the activity file. How to do that? Ignore the first column returned from read.table, and the 2nd column is the descriptive activity names for activities, then index them with data$activity directly. Great right? After this sentence, activities will be replaced with descriptive activity names.

Goal 2: Extracts only the measurements on the mean and standard deviation for each measurement.

Then extract required data for tidy data. But I have to add two columns, 'subject' and 'activity'. I can create new variable to save them, but keep them in data frame will be nice. And it short the code, I don't have to type more letters later. Grep will return the index for matched items, so it's great, I can use it directly. And I found only mean() or std() have meaning, if you use mean as pattern, it will include some unwanted data, like meanfeq... So I use mean\( as pattern.

Goal 5: Creates a second, independent tidy data set with the average of each variable for each activity and each subject.

Calculate the mean for each activity and each subject. All datas are ready to use, what you need is just a beautiful formula. Goal~~~ I love aggregate, I love formula too.

Last step, save the data.

Why I want to write a short code for this project?

Short means you can read it faster, because it has less letters. You don't need to remember too much things (variables, processes...), because it's short right? Normally, short means simple, so I hope this can help you to understand what I want to do easier. And the code will looks beautiful as well, right?

And the most important reason: I'm too lazy to write so much code~~~ LOL