1. Video recordings:

	The data collected during the user study were screen video recordings of the participant's browsing
	The data is overlayed with colored visualization of participant's visual attention fixations at a given instance of time


2. Video recording title:

	Each recording is titled as "sXXX" where s stands for subject and followed by their numeric id
	eg. Video recording titled "s022" is the recording of participant ID 22


3. Annotations:

	There are two annotations created namely
	2017_Spring_YouTube-Attention_Session1-24_HomePage_Chaitra
	2017_Spring_YouTube-Attention_Session1-24_TimeSpentHomePage_Chaitra

	Both these files consist of 3 common columns namely
	Participant ID, Trial and Seconds

	Each trial for a given participants represents the time when the YouTube homepage loads 
	till the time a video is selected to watch


	The "2017_Spring_YouTube-Attention_Session1-24_TimeSpentHomePage_Chaitra" file provides 
	the amount of time spent on the homepage searching for a video to watch in seconds
	Start time is noted when the YouTube homepage loads completely
	End time is noted when a video selection is made
	Time Spent is computed based on the start and end timestamps of the video recording 
	for each trial for a participant


	The "2017_Spring_YouTube-Attention_Session1-24_HomePage_Chaitra" consists of numerical statistics of
	the following areas of interests:

	Watched Ad: Indicated by '1' when participant watched the complete ad after it loaded
	Thumbnail: Number of thumbnails viewed before a video is selected
	Thumbnail After Selection: Number of thumbnails viewed after a video was selected
	Repeated Thumbnail: Number of thumbnails viewed more than once in a trial
	Titled Text: Number of title text viewed before a video is selected
	Titled Text After Selection: Number of titled text viewed after a video was selected
	Repeated Titled Text: Number of title text looked at more than once in a trial
	Thumbnail seen before: Number of times the thumbnail was viewed before title text 
	Titled Text seen before: Number of times title text was viewed before thumbnai
	Both viewed: Number of times both the thumbnail and title text of a video was viewed
	Shelf Titles: Number of participant viewed the title of a self
	Row: Indicated by '1' when participant looked at the titled text only in a given shelf
	Load more: Number of times partitipants clicked on load more button at the bottom of the homepage
	No. of Subcribers: Number of times partitipants looked at the number of subcribers for a channel

4. R-Scripts:


	Three scripts were used to analyze the data and to generate all the plots:
	1. youtube-analysis-homepage
	2. youtube-analysis-homepage_2
	3. youtube-analysis-timespent

	The order in which you run the scripts doesn't matter

	1. youtube-analysis-homepage: 

	This script takes in "2017_Spring_YouTube-Attention_Session12345678_ThumbnailsAndTitleText_Chaitra.csv" as input
	a. First, we have defined a function to draw error bars in barplots.
	b. We then computed the number of thumnails viewed by each participant by using the aggregate function.
	c. Next, after finding the mean and sd of the thumbnails viewed by each participant, we generated a scattered plot for the same. 
	d. After computing the average and sd of the number of thumbnails viewed per trial by each participant was computed and a bar plot was generated using the error bars.
	e. steps b to d were repeated for the other dependent variables: TitleTextViewed, ThumbnailsFirst and TitleTextFirst
	f. Figure 4c, 4d and 4e were generated using the boxplot and barplot commands
	g. Two One-tailed Paired t-tests were run. 

	2. youtube-analysis-homepage_2

	This script takes in "2017_Spring_YouTube-Attention_Session12345678_ThumbnailsAndTitleText_Chaitra.csv" as input
	It calculates the number of thumbnails viewed by a participant per second with mean and sd.

	3. youtube-analysis-timespent

	This script takes in "TimeSpentHomePage.csv" as input
	a. Timespent by participants across all trials was` computed across all trials along with mean and sd
	b. A scattered plot for the same was generated (used as figure 4b)
	c. Function to display the error bars was created
	d. Next, bar plot of time spent on homepage per trial vs Participant ID was generated (used as figure 4a)
	e. Percentage of time spent homepage was computed and a scattered plot for the same was generated
	f. A function to convert mm:ss to seconds was created
	g. Finally, a histogram for length of videos watched was generated