Twitter user analysis

Analyzing Social Media Data in R

Vivek Vijayaraghavan

Data Science Coach

Lesson Overview

  • friends_count and followers_count of a user
  • Interpret golden ratio for brand promotion
  • Twitter lists to identify users interested in a product
Analyzing Social Media Data in R

Followers vs friends

Twitter friends and followers

  • Followers are users following a twitter user
  • Friends are people a specific twitter user is following
Analyzing Social Media Data in R

Twitter follower vs following ratio

Golden ratio formula

  • Used by marketers to strategize promotions
Analyzing Social Media Data in R

Positive and negative ratios

  • Positive ratio: more followers than friends for a user
  • Negative ratio: more friends than followers for a user
Analyzing Social Media Data in R

Extract user information

# Search for 1000 tweets on #fitness
tweet_fit <- search_tweets("#fitness", n = 1000)

# Extract user information
user_fit <- users_data(tweet_fit)
Analyzing Social Media Data in R

Extract user information

# View column names of the user data
names(user_fit)

column names of user data

Analyzing Social Media Data in R

Extracting followers_count and friends_count

  • Aggregate user screen names against followers and friends counts
# Aggregate screen_name, followers_count & friends_count
library(dplyr)
counts_df <- user_fit %>%
               group_by(screen_name) %>%
               summarize(follower = mean(followers_count),
               friend = mean(friends_count))
Analyzing Social Media Data in R

Extracting followers_count and friends_count

head(counts_df)

screen_name      follower     friend
<chr>              <dbl>       <dbl>
__seokjinnie124     209        454        
_Aminata            623        523        
_amsvn              167        126        
_arweeennn          539        801        
_asof_             1336        455        
_blendac            833        195
Analyzing Social Media Data in R

The golden ratio

# Create a column to calculate the golden ratio
counts_df$ratio <- follow_df$follower/follow_df$friend
head(counts_df$ratio)

[1] 0.4603524 1.1912046 1.3253968 0.6729089 2.9362637 4.2717949
Analyzing Social Media Data in R

Explore users based on the ratio

  • Examine golden ratios to understand user types
# Sort the data frame in decreasing order of follower count
counts_sort <- arrange(counts_df, desc(follower))
Analyzing Social Media Data in R

Explore users based on the ratio

# Select rows where the follower count is greater than 30000
counts_sort[counts_sort$follower>30000,]

screen_name    follower    friend    ratio        
<chr>            <dbl>      <dbl>    <dbl>
mashable         9817699    2783     3528    
MensHealthMag    4528421    1111     4076    
Sophie_Choudry   2367827    157      15082    
thewebmaster_    103936     6508     16    
qwikad           92932      89557    1    
Rharvley         90464      19484    5    
SayWhenLA        68122      6680     10
  • Medium to promote products on fitness
Analyzing Social Media Data in R

Explore users based on ratio

# Select rows where the follower count is less than 2000
counts_sort[counts_sort$follower<2000,]

screen_name    follower     friend    ratio        
<chr>            <dbl>       <dbl>    <dbl>
workout_ehime    1960        1027       2    
SardImperium     1932        256        8    
Deem_Hoops       1912        1520       1    
kaykay_inem      1890        443        4    
bhealhty         1855        3066       1
  • Position adverts on individual accounts for targeted promotion
Analyzing Social Media Data in R

User analysis with twitter lists

  • Twitter list is a curated group of twitter accounts
  • Twitter users subscribe to lists of interest
Analyzing Social Media Data in R

Extract lists subscribed to

# Get all lists "Playstation" subscribes to
lst_playstation <- lists_users("PlayStation")
lst_playstation[,1:4]

list_id          name                      uri               subscriber_count
<chr>            <chr>                    <chr>                      <int>
58505230       PS Family         /PlayStation/lists/ps-family         136    
4747423        GameDevelopers    /PlayStation/lists/gamedevelopers    467    
2490894        gaming            /PlayStation/lists/gaming            658
Analyzing Social Media Data in R

Extract subscribers to a list

# Extract 100 subscribers of the "gaming" list owned by "Playstation"
list_PS_sub <- lists_subscribers(slug = "gaming", owner_user = "PlayStation", n = 100)
Analyzing Social Media Data in R

View screen names of subscribers

# View screen names of the subscribers
list_PS_sub$screen_name

Screen names of subscribers

Analyzing Social Media Data in R

User information of list subscribers

# Create a list of four screen names
users <- c("Morten83032201","ndugumr", "WOLF210_Warrior", "souransb")

# Extract user information
users_PS_gaming <- lookup_users(users)

user_id                   status_id              created_at          screen_name
<chr>                       <chr>               <S3: POSIXct>          <chr>
1158299850573791233    1172604921121824769    2019-09-13 20:16:13    Morten83032201    
894525207620321280     1183293767215992832    2019-10-13 08:09:53    ndugumr    
325760816              1182867378293616640    2019-10-12 03:55:34    WOLF210_Warrior    
469270931              511997829384904704     2014-09-16 21:59:29    souransb
Analyzing Social Media Data in R

Let's practice!

Analyzing Social Media Data in R

Preparing Video For Download...