Lab 09
Amelia Jerkatis
Task#1: Data Exploration - Tableau
1A) Generally describe the data (summary)
mydata = read.csv(file="data/march_madness.csv")
mydatasummary(mydata)## msg_id
## Min. :8.522e+17
## 1st Qu.:9.778e+17
## Median :9.779e+17
## Mean :9.785e+17
## 3rd Qu.:9.798e+17
## Max. :9.816e+17
##
## avatar
## https://pbs.twimg.com/profile_images/965330916007608320/PIIJeigi_bigger.jpg : 56
## https://abs.twimg.com/sticky/default_profile_images/default_profile_bigger.png: 41
## https://pbs.twimg.com/profile_images/585091148965810177/CzafKE6d_bigger.jpg : 32
## https://pbs.twimg.com/profile_images/1422356936/square_MICH_bigger.jpg : 25
## https://pbs.twimg.com/profile_images/884405015741444096/fUiezbHX_bigger.jpg : 21
## https://pbs.twimg.com/profile_images/773277131430297600/1OgEyc-m_bigger.jpg : 11
## (Other) :2344
## verified username fullname
## Min. :0.0000 @umichbball : 56 Michigan Men's Basketball: 56
## 1st Qu.:0.0000 @RamblersMBB : 32 Loyola Basketball : 32
## Median :0.0000 @MichiganOnBTN : 25 Michigan On BTN : 25
## Mean :0.1842 @marchmadness : 21 NCAA March Madness : 21
## 3rd Qu.:0.0000 @CharlieDeMar : 11 Charlie De Mar : 11
## Max. :1.0000 @VoteForDoctorY: 11 Doctor Y : 11
## (Other) :2374 (Other) :2374
## tweet
## \n Go Blue\n : 3
## \n Go Blue!\n : 3
## \n I hope you keep shinning. You Guys seem humble, you work hard, you're having fun, you play as a Team and you're inspired by #SisterJean who's dope!!! I'm rooting 4 you #LoyolaChicago #marchmadness #FinalFour #Loyola #NCAA #MarchMadness2018 #Basketball #OnwardLU #NikeBasketball\n: 3
## \n Looks like #SisterJean didnâ\u0080\u0099t pray hard enough! You let your team down Jean. Congrats @umichbball! See you Monday night! #MarchMadness2018 #GoBlue\n : 3
## \n This is fantastic! Prayers are being answered! Plus the team has Incredible talent! Looking forward to the next game ! #SisterJean #Ramblers #FinalFour #MarchMadness2018\n : 3
## \n .@nikolasmarty just wondering if you're a #MarchMadness2018 fan???? Lol I'm sure just as happy as Iam that @umichbball is in the #FinalFour #GoBlue #WhereInTheWorldIsOSUBasketball #Wolverines\n : 2
## (Other) :2513
## time reply retweets
## 2018-03-25T04:01:31Z: 3 Min. : 0.000 Min. : 0.00
## 2018-03-25T04:03:48Z: 3 1st Qu.: 0.000 1st Qu.: 0.00
## 2018-03-25T04:44:12Z: 3 Median : 0.000 Median : 0.00
## 2018-03-25T06:57:16Z: 3 Mean : 3.109 Mean : 41.67
## 2018-03-25T08:14:33Z: 3 3rd Qu.: 1.000 3rd Qu.: 6.00
## 2018-03-25T10:10:18Z: 3 Max. :592.000 Max. :9328.00
## (Other) :2512
## favorite links
## Min. : 0.0 @umichbball : 68
## 1st Qu.: 0.0 @RamblersMBB : 54
## Median : 2.0 #MarchMadness2018 : 53
## Mean : 198.7 #SisterJean : 52
## 3rd Qu.: 26.0 @umichbball; #GoBlue: 22
## Max. :32149.0 (Other) :1866
## NA's : 415The data consists of 11 columns with 2530 records. The first column is a unique identifier that is the message id. The second column is a link to the avatar of the tweeter. The verified column identifies if the account that tweeted was verified. There are a few verified accounts in this dataset, but not very many since the mean of the verified column is 0.1842. The username identifies the username of the tweeter and the fullname column identifies the tweeters actual name. These are both text columns. The next column is tweet, which has the text of the tweet. The time column is a timestamp of when the tweet was posted. The reply, retweet, and favorite columns identifies the number of times the tweet was replied to, retweeted, or favorited respectively. The tweet with the highest replys had 592 replys, the tweet with the highest retweets had 9328 retweets, and the tweet with the most favorites had 32149 favorites. The links column identifies if there were any links attached to the tweet. This can be usernames, hyperlinks, or hashtags. Umichbball and RamblersMBB were the highest mentioned accounts.
1B) Use tableau to create at least 5 plots
1C) Explain each plot make a relation to date of the tweets/time
knitr::include_graphics('imgs/Top Tweeters.jpg')
This graph looks at the top tweeters during the March Madness games. The University of Michigan basketball account and the Loyola Basketball account are the two accounts that tweeted the most during the games. There were several other accounts that tweeted multiple times throughout the tournament, but the majority of the accounts only tweeted once or twice.
knitr::include_graphics('imgs/Top Tweeters Bubble Chart.jpg')
This is another graph that shows the top tweeters during the tournament, however this is a bubble graph instead of a bar graph. This graph is also filtered to accounts that tweeted at least 5 times. This makes the graph more easily viewable.If this graph was being viewed in Tableau, you would be able to hover over the bubbles without labels in order to identify which accounts those are.
knitr::include_graphics('imgs/Time of Tweets.jpg')
This graph looks at the number of tweets per minute during the tournament. The most tweets occured during the games, and people did not tweet as much about March Madness or these teams when the games were not occurring.
knitr::include_graphics('imgs/Hashtag Frequency.jpg')
This is a bubble chart of the frequency of the hashtags used during the tournament. This graph was filtered so that hashtags that were only used a few times were not included. The top hashtags used can be identified by the size of the bubble. #sister jean, #finalfour, #marchmadness2018, #goblue, #marchmadness, and #onwardlu were the top hashtags used. Once again, if this graph was being viewed in Tableau, the labels for the other bubbles could be seen when you hover over the bubble.
knitr::include_graphics('imgs/Favorite vs. Reply vs. Retweets.jpg')
This graph looks at the number of favorites, replies, and retweets plotted against time. You can tell from this graph that tweets were favorited more than replied to or retweeted. You can also see that the most tweets were favorited during the end of the championship game on April 1st.
Task#2: Data Analysis
2A) Based on your plots and data description make give a general narrative for the image of loyola in twitter
Loyola was active on twitter during the tournament, however not as active as Michigan. I think the case for this difference is that Michigan played in the championship game on April 1st, whereas Loyola had already finished their tournament run. Even though this is the case, #sisterjean was still the most used hashtag, which shows just how much she captured America’s heart during the tournament.
2B) Use descriptive statistics to backup your arguments
summary(mydata)## msg_id
## Min. :8.522e+17
## 1st Qu.:9.778e+17
## Median :9.779e+17
## Mean :9.785e+17
## 3rd Qu.:9.798e+17
## Max. :9.816e+17
##
## avatar
## https://pbs.twimg.com/profile_images/965330916007608320/PIIJeigi_bigger.jpg : 56
## https://abs.twimg.com/sticky/default_profile_images/default_profile_bigger.png: 41
## https://pbs.twimg.com/profile_images/585091148965810177/CzafKE6d_bigger.jpg : 32
## https://pbs.twimg.com/profile_images/1422356936/square_MICH_bigger.jpg : 25
## https://pbs.twimg.com/profile_images/884405015741444096/fUiezbHX_bigger.jpg : 21
## https://pbs.twimg.com/profile_images/773277131430297600/1OgEyc-m_bigger.jpg : 11
## (Other) :2344
## verified username fullname
## Min. :0.0000 @umichbball : 56 Michigan Men's Basketball: 56
## 1st Qu.:0.0000 @RamblersMBB : 32 Loyola Basketball : 32
## Median :0.0000 @MichiganOnBTN : 25 Michigan On BTN : 25
## Mean :0.1842 @marchmadness : 21 NCAA March Madness : 21
## 3rd Qu.:0.0000 @CharlieDeMar : 11 Charlie De Mar : 11
## Max. :1.0000 @VoteForDoctorY: 11 Doctor Y : 11
## (Other) :2374 (Other) :2374
## tweet
## \n Go Blue\n : 3
## \n Go Blue!\n : 3
## \n I hope you keep shinning. You Guys seem humble, you work hard, you're having fun, you play as a Team and you're inspired by #SisterJean who's dope!!! I'm rooting 4 you #LoyolaChicago #marchmadness #FinalFour #Loyola #NCAA #MarchMadness2018 #Basketball #OnwardLU #NikeBasketball\n: 3
## \n Looks like #SisterJean didnâ\u0080\u0099t pray hard enough! You let your team down Jean. Congrats @umichbball! See you Monday night! #MarchMadness2018 #GoBlue\n : 3
## \n This is fantastic! Prayers are being answered! Plus the team has Incredible talent! Looking forward to the next game ! #SisterJean #Ramblers #FinalFour #MarchMadness2018\n : 3
## \n .@nikolasmarty just wondering if you're a #MarchMadness2018 fan???? Lol I'm sure just as happy as Iam that @umichbball is in the #FinalFour #GoBlue #WhereInTheWorldIsOSUBasketball #Wolverines\n : 2
## (Other) :2513
## time reply retweets
## 2018-03-25T04:01:31Z: 3 Min. : 0.000 Min. : 0.00
## 2018-03-25T04:03:48Z: 3 1st Qu.: 0.000 1st Qu.: 0.00
## 2018-03-25T04:44:12Z: 3 Median : 0.000 Median : 0.00
## 2018-03-25T06:57:16Z: 3 Mean : 3.109 Mean : 41.67
## 2018-03-25T08:14:33Z: 3 3rd Qu.: 1.000 3rd Qu.: 6.00
## 2018-03-25T10:10:18Z: 3 Max. :592.000 Max. :9328.00
## (Other) :2512
## favorite links
## Min. : 0.0 @umichbball : 68
## 1st Qu.: 0.0 @RamblersMBB : 54
## Median : 2.0 #MarchMadness2018 : 53
## Mean : 198.7 #SisterJean : 52
## 3rd Qu.: 26.0 @umichbball; #GoBlue: 22
## Max. :32149.0 (Other) :1866
## NA's : 415Here you can see that @umichbball had 56 tweets whereas @RamblersMBB only had 32. Also, @umichbball was linked to (on its own) 68 times whereas @RamblersMBB was linked to (on its own) 54 times. This shows that Michigan both utilized twitter more and was talked about more on twitter throughout the tournament. But, as mentioned previously I suspect this is because Michigan played in the championship game and Loyola did not.
2C) Any recommendations to Loyola’s marketing team
I would recommend that Loyola be more active on twitter during the games, and to keep that activity going throughout the entire tournament. The more active that Loyola’s accounts are, the more they will be retweeted, favorited, and replied to, and the more people will be mentioning them in their tweets.
Task#3: Watson Analytics
3A) Use watson analytics to explore the data
3B) Give at least 3 plots or discoveries using watson. Explain your findings.
knitr::include_graphics('imgs/Top Retweeted.png')
This graph looks at the usernames that had the most retweets. The University of Michigan Basketball account had the most retweets with the Loyola Basketball account and the March Madness accounts coming in second and third.
knitr::include_graphics('imgs/Top Favorited.png')
This graph looks at the usernames with the tweets that were favorited the most. This graph is interesting because the accounts are different than the top retweeted accounts. The usernames with the most favorited tweets were the Bleacher Report account and the University of Michigan Football account. I would be interested to look and see if each of these had one specific tweet that was favorited alot, or if they had consistently good tweets that were favorited an average amount.
knitr::include_graphics('imgs/Top Replied.png')
This graph shows the usernames with the most replies to their tweets. The top three are similar to the accounts with the most retweets, however the Loyola Basketball account has a lot more replies than the March Madness account, whereas they were close in number for number of retweets.