Homework_02

pdf

School

Yale University *

*We aren’t endorsed by this school

Course

123B

Subject

Statistics

Date

Feb 20, 2024

Type

pdf

Pages

Uploaded by JusticeRoseLion35

Question 1.1 Homework 02 Data Handling, Graphics, More R Due by 11:59pm, Friday, 2.2.24 S&DS 230/530/ENV 757 (1) Obama Tweets: Retweets vs. Favorites A .CSV file containing recent Tweets from former President Barack Obama can be downloaded HERE . The data is sorted by date, most recent at the top. The variables (columns) are: • text : the body of the tweet • date : when the tweet was sent, original format • date2 : when the tweet was sent, JUST the day (not the time of day) • retweet_count : how many people retweeted this tweet • favorite_count : how many people favorited this tweet • is_retweet : whether or not this tweet is a retweet of someone else’s tweet • source : device used to send the tweet • is_quote : is the tweet a quote of someone else • is_reply : is the tweet a reply There are two ways in which other Twitter users can indicate support for a tweet: favoriting and retweeting . For example, if a tweet has favorite_count = 5 and retweet_count = 10, then this suggests that 5 people favorited the tweet (saved it) and 10 people retweeted it (broadcasted it to their followers). (1.1) Insert an R code chunk right below this that imports the data into a dataframe called recent . Note that the data is sorted in reverse time order. Get the header names of recent to confirm that the data imported correctly. Look at the first few rows of the data and the final few rows of the data. Also get the dimension of recent . What is the date range of the tweets? How many tweets does this dataset include? recent <- read.csv ( "http://reuningscherer.net/S&DS230/data/ObamaTweets.csv" , header = TRUE ) names (recent) ## [1] "X" "text" "date" "source" ## [5] "is_quote" "is_retweet" "is_reply" "favorite_count" ## [9] "retweet_count" "date2" head (recent) ## X ## 1 1 ## 2 2

Question 1.1 ## 3 3 ## 4 4 ## 5 5 ## 6 6 ## text ## 1 If you need to get covered, head over to https://t.co/DEjP7PTKgn. Open enrollment ends this Sunday, January 15. ## 2 I’ll always be proud to have my name associated with a program that’s helping folks get the health care they need. https://t.co/mnOTx4jdOA ## 3 This week, Illinois joined states across the country in passing a historic gun violence prevention package. This would not have been possible without the work of grassroots organizers, survivors and their families, and lawmakers who channeled their anguish into bold action. ## 4 RT @welcomeus: Through new humanitarian sponsorship programs, vulnerable newcomers can now find safety in the U.S. with the support of a sp… ## 5 RT @POTUS: Nearly 16 million people have signed up for coverage through the Affordable Care Act since the start of open enrollment. That’s… ## 6 If you haven't already, I hope you’ll take some time to watch Descendant on @Netflix. It's an important documentary that shares the story of the descendants of the Clotilda, the last known ship carrying enslaved folks from Africa, as they uncover the truth about the past. https://t.co/GQGDoGvwKQ ## date ## 1 2023-01-13 13:30:43 ## 2 2023-01-13 13:30:43 ## 3 2023-01-12 08:30:25 ## 4 2023-01-11 10:45:56 ## 5 2023-01-11 09:31:33 ## 6 2023-01-10 14:37:04 ## source ## 1 <a href="http://twitter.com/download/iphone" rel="nofollow">Twitter for iPhone</a> ## 2 <a href="http://twitter.com/download/iphone" rel="nofollow">Twitter for iPhone</a> ## 3 <a href="http://twitter.com/download/iphone" rel="nofollow">Twitter for iPhone</a> ## 4 <a href="http://twitter.com/download/iphone" rel="nofollow">Twitter for iPhone</a> ## 5 <a href="http://twitter.com/download/iphone" rel="nofollow">Twitter for iPhone</a> ## 6 <a href="https://mobile.twitter.com" rel="nofollow">Twitter Web App</a> ## is_quote is_retweet is_reply favorite_count retweet_count date2 ## 1 FALSE FALSE TRUE 4045 847 2023-01-13 ## 2 FALSE FALSE FALSE 15256 1563 2023-01-13

Question 1.1 ## 3 FALSE FALSE FALSE 28154 3760 2023-01-12 ## 4 FALSE FALSE FALSE 0 347 2023-01-11 ## 5 FALSE FALSE FALSE 0 3145 2023-01-11 ## 6 FALSE FALSE TRUE 8404 1310 2023-01-10 tail (recent) ## X ## 1995 1995 ## 1996 1996 ## 1997 1997 ## 1998 1998 ## 1999 1999 ## 2000 2000 ## text ## 1995 It's time for the United States to #LeadOnLeave — show your support if you agree: https://t.co/XtTmJH00Y9 ## 1996 Retweet if you believe it's time for the United States to #LeadOnLeave. https://t.co/8jsAkBBqzx ## 1997 Speak up for a fair hearing for Judge Merrick Garland: https://t.co/Tf8yVdukBx #DoYourJob ## 1998 This is unprecedented. #DoYourJob https://t.co/kMdn564dCW ## 1999 Add a comment if you agree: American workers shouldn't have to choose between their health and a paycheck. https://t.co/AbSVkV1sEr ## 2000 Working families in America should have the basic security of paid sick leave. #LeadOnLeave https://t.co/0jxbEnUUJg ## date ## 1995 2016-04-11 10:11:20 ## 1996 2016-04-11 08:34:06 ## 1997 2016-04-08 14:23:02 ## 1998 2016-04-08 11:52:17 ## 1999 2016-04-08 10:04:33 ## 2000 2016-04-08 08:45:49 ## source ## 1995 <a href="http://twitter.com" rel="nofollow">Twitter Web Client</a> ## 1996 <a href="http://twitter.com" rel="nofollow">Twitter Web Client</a> ## 1997 <a href="http://twitter.com" rel="nofollow">Twitter Web Client</a> ## 1998 <a href="http://twitter.com" rel="nofollow">Twitter Web Client</a> ## 1999 <a href="http://twitter.com" rel="nofollow">Twitter Web Client</a> ## 2000 <a href="http://twitter.com" rel="nofollow">Twitter Web Client</a> ## is_quote is_retweet is_reply favorite_count retweet_count date2 ## 1995 FALSE FALSE FALSE 1544 378 2016-04-11 ## 1996 FALSE FALSE FALSE 6015 3184 2016-04-11 ## 1997 FALSE FALSE FALSE 2271 762 2016-04-08 ## 1998 FALSE FALSE FALSE 4388 1890 2016-04-08 ## 1999 FALSE FALSE FALSE 3141 724 2016-04-08 ## 2000 FALSE FALSE FALSE 7082 1732 2016-04-08

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version

Access to all documents
Unlimited textbook solutions
24/7 expert homework help

Question 1.1, 1.2, 1.3 dim (recent) ## [1] 2000 10 The tweets in this data frame range from 2016/04/08 to 2023/01/13. There are 2000 tweets included in this data frame (1.2) Create a table that shows how many of the Tweets were quotes (that is, President Obama retweeted someone elses tweet but added additional commentary), and call this object table1 . Show the results of table1 . Write a single line that calculates the percent of Tweets that were quotes, rounds this value to two decimal places, multipies the results by 100, and pastes on a “%” symbol. There should be no space between the number and the ‘%’ symbol. Finally, have the entire line read “?% of Obama’s tweets were quotes”, where ? is the calculated percentage. table1 <- table (recent $ is_quote) table1 ## ## FALSE TRUE ## 1817 183 paste ( round ((table1[ 2 ] / (table1[ 1 ] + table1[ 2 ])), 2 ) * 100 , "% of Obama's tweets were quotes" , sep = "" ) ## [1] "9% of Obama's tweets were quotes" (1.3) Get summary statistics for both favorite_count and retweet_count . Make histograms for each of these two variables as well. Put a title on each histogram, label the horizontal axis, and make the bars orange. How would you describe the shape of these distributions (use words like ‘symmetric’ or ‘skewed’, or perhaps the name of some distribution that has a similar shape . . .)? summary (recent $ favorite_count) ## Min. 1st Qu. Median Mean 3rd Qu. Max. ## 0 3969 15358 96327 78010 4010967 summary (recent $ retweet_count) ## Min. 1st Qu. Median Mean 3rd Qu. Max. ## 198 1164 3071 16841 12153 1435375 hist (recent $ favorite_count, col = "orange" , xlab = "Number of favorites" , main = "Histogram of favorited tweets by Obama" )

Question 1.3 hist (recent $ retweet_count, col = "orange" , xlab = "Number of retweets" , main = "Histogram of retweets by Obama" )

Question 1.3 Both graphs display a right skewed unimodal distribution and resemble exponential distributions (1.4) Get summary statistics for favorite_count FIRST for the observations for which is_quote is TRUE , then for the observations for which ’is_quote is FALSE`. Compare the medians of these two distributions - what do you observe? summary ( subset (recent, is_quote == TRUE , select = favorite_count)) ## favorite_count ## Min. : 0 ## 1st Qu.: 14492 ## Median : 58453 ## Mean : 114081 ## 3rd Qu.: 164038 ## Max. :1159695 summary ( subset (recent, is_quote == FALSE , select = favorite_count)) ## favorite_count ## Min. : 0 ## 1st Qu.: 3474 ## Median : 13278 ## Mean : 94539

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version

Access to all documents
Unlimited textbook solutions
24/7 expert homework help

Question 1.4, 1.5, 1.6, 1.7 ## 3rd Qu.: 69172 ## Max. :4010967 The median tweet that was quoted gets favorited more than the median tweet that are not quoted. (1.5) Create a new dataframe called recent_NoQuote that contains all data from recent for which is_quote is FALSE (essentially, we’re removing quotes and only looking at strictly original texts). USE THIS NEW DATAFRAME for the remainder of this problem set. Get the dimension of this dataframe to make sure the remaining number of rows (and columns) are consistent with the results in part 1.2. recent_NoQuote <- subset (recent, is_quote == FALSE ) dim (recent_NoQuote) ## [1] 1817 10 (1.6) Make two new variables as a part of recent_NoQuote which will be the log base 10 transformations of favorite_count and retweet_count . Call these variables log10favCnt and log10reCnt , respectively. The function you want to take log base 10 is called log10() . Note - you can add a variable to dateframe by simply creating a name using the $ operator and then assigning it the desired value : e.g. recent_NoQuote$log10facCnt <- (whatever you want to assign this) recent_NoQuote $ log10favCnt <- log10 (recent_NoQuote $ favorite_count) recent_NoQuote $ log10reCnt <- log10 (recent_NoQuote $ retweet_count) (1.7) Make histograms of these two new log-scale variables. Put a title on each histogram, label the horizontal axis, and make the bars orange. How would you describe the shape of these transformed distributions (use words like ‘symmetric’ or ‘skewed’, or perhaps the name of some distribution that has a simlar shape . . .)? hist (recent_NoQuote $ log10favCnt, col = "orange" , xlab = "Number of favorites (log_10)" , main = "Histogram of Favorited Tweets by Obama (log scale)"

Question 1.7 hist (recent_NoQuote $ log10reCnt, col = "orange" , xlab = "Number of favorites (log_10)" , main = "Histogram of Retweeted Tweets by Obama (log scale)" )

Question 1.7 These histograms are still right skewed unimodal distribution but look more symmetric (1.8) Make a plot of the number of times that each tweet was favorited vs. the number of times a tweet was retweeted. Put favorite_count on the x-axis and retweet_count on the y-axis. Label your axes, put on a main title, and make the plot characters blue. plot ( x = recent_NoQuote $ favorite_count, y = recent_NoQuote $ retweet_count, main = "Favorite vs retweet counts of tweets by Obama" , ylab = "number of times retweeted" , xlab = "number of times favorited" , col = "blue"

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version

Access to all documents
Unlimited textbook solutions
24/7 expert homework help

Question 1.8, 1.9 (1.9) Repeat part (1.8) but use the log-transformed variables. Label your axes, put on a main title, and make the plot characters red. How does the scatterplot on the log-scale compare to the scatterplot on the raw scale? Which one do you prefer? plot ( x = recent_NoQuote $ log10favCnt, y = recent_NoQuote $ log10reCnt, main = "Favorite vs retweet counts of tweets by Obama (log scale)" , ylab = "number of times retweeted (log_10)" , xlab = "number of times favorited (log_10)" , col = "red

Question 1.9, 1.10 I prefer the log scale plot because it shows a much more clear linear trend between number of times favorited and number of times retweeted and the points is scaled towards the center of the plot (1.10) Create two new variables on the recent_NoQuote dataframe called year and month that will contain respectively the year and month the tweet was created. You’ll need to look up how to use the function substr() . You’ll also need to use the as.numeric() function to make sure that both new variables are numbers. Show the first 20 observations for each resulting variable. recent_NoQuote $ year <- as.numeric ( substr (recent_NoQuote $ date2, 1 , 4 )) recent_NoQuote $ year[ c ( 1 : 20 )] ## [1] 2023 2023 2023 2023 2023 2023 2023 2023 2023 2023 2023 2022 2022 2022 2022 ## [16] 2022 2022 2022 2022 2022 recent_NoQuote $ month <- as.numeric ( substr (recent_NoQuote $ date2, 6 , 7 )) recent_NoQuote $ month[ c ( 1 : 20 )] ## [1] 1 1 1 1 1 1 1 1 1 1 1 12 12 12 12 12 12 12 12 12

Question 1.11 (1.11) Repeat part (1.9) BUT only for 2018 and 2022 First, create a dataframe called recent_3 that only has observations from the specified years. You might want to use the %in% operator on your newly created variable year. Use this new dataframe to make your plot. Use the graphics option pch = 19 to get solid round points, and use the graphics option col = as.factor(year) to make different colors for 2018 and 2022 The final line of code below will add a legend to the top left of the plot. #put R code here to make dataframe and plot recent_3 <- subset (recent_NoQuote, year %in% c ( "2022" , "2018" )) plot ( x = recent_3 $ log10favCnt, y = recent_3 $ log10reCnt, main = "Favorite vs retweet counts of tweets by Obama in 2018 and 2022(log scale)" , ylab = "number of times retweeted (log_10)" , xlab = "number of times favorited (log_10)" , col = as.factor (recent_3 $ year), pch= 19 ) legend ( "topleft" , legend = c ( "2018" , "2022" ), col = c ( 1 , 2 ), pch = 19 ) (1.12) Write no more than three sentences that describe what you see. Does the pattern appear any different between 2018 and 2022

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version

Access to all documents
Unlimited textbook solutions
24/7 expert homework help

Question 1.12 Both the 2018 and 2022 data display linear and positively sloped relationship. The 2018 data is clustered more towards the top right while the 2022 data is more clustered towards the bottom left, showing that the tweets in 2018 were retweeted and favorited more than those in 2020

Homework_02

Related Documents