The #NGSSTweets Twitter Social Network

Research Question

Do the actors with high in-degree centrality also tend to have high out-degree centrality?
Which actors serve as transmitters, transceivers and transcenders in the NGSS tweets network?

Loading Libraries and Data

library(tidyverse)

## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ dplyr     1.1.0     ✔ readr     2.1.4
## ✔ forcats   1.0.0     ✔ stringr   1.5.0
## ✔ ggplot2   3.4.1     ✔ tibble    3.1.8
## ✔ lubridate 1.9.2     ✔ tidyr     1.3.0
## ✔ purrr     1.0.1     
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag()    masks stats::lag()
## ℹ Use the ]8;;http://conflicted.r-lib.org/conflicted package]8;; to force all conflicts to become errors

library(igraph)

## Warning: package 'igraph' was built under R version 4.2.3

## 
## Attaching package: 'igraph'
## 
## The following objects are masked from 'package:lubridate':
## 
##     %--%, union
## 
## The following objects are masked from 'package:dplyr':
## 
##     as_data_frame, groups, union
## 
## The following objects are masked from 'package:purrr':
## 
##     compose, simplify
## 
## The following object is masked from 'package:tidyr':
## 
##     crossing
## 
## The following object is masked from 'package:tibble':
## 
##     as_data_frame
## 
## The following objects are masked from 'package:stats':
## 
##     decompose, spectrum
## 
## The following object is masked from 'package:base':
## 
##     union

library(tidygraph)

## Warning: package 'tidygraph' was built under R version 4.2.3

## 
## Attaching package: 'tidygraph'
## 
## The following object is masked from 'package:igraph':
## 
##     groups
## 
## The following object is masked from 'package:stats':
## 
##     filter

library(ggraph)

## Warning: package 'ggraph' was built under R version 4.2.3

library(skimr)

## Warning: package 'skimr' was built under R version 4.2.3

library(dplyr)
library(readxl)
library(tidytext)
library(dplyr)

ngss_tweets <- read_xlsx("ngss_tweets.xlsx")

ngss_tweets

## # A tibble: 402 × 91
##    user_id      statu…¹ created_at          scree…² text  source displ…³ reply…⁴
##    <chr>        <chr>   <dttm>              <chr>   <chr> <chr>    <dbl> <chr>  
##  1 316633205    136571… 2021-02-27 17:33:27 loyr26… "Swi… Twitt…     117 <NA>   
##  2 316633205    136321… 2021-02-20 20:02:37 loyr26… "Was… Twitt…     200 <NA>   
##  3 13391550230… 136570… 2021-02-27 17:04:18 Ngss_18 "Ras… Twitt…      34 <NA>   
##  4 13391550230… 136529… 2021-02-26 13:51:35 Ngss_18 "Kan… Twitt…      28 <NA>   
##  5 13391550230… 136565… 2021-02-27 13:44:20 Ngss_18 "Seb… Twitt…      44 <NA>   
##  6 13391550230… 136529… 2021-02-26 13:45:00 Ngss_18 "Bin… Twitt…      71 <NA>   
##  7 13391550230… 136549… 2021-02-27 02:52:34 Ngss_18 "Buk… Twitt…     194 <NA>   
##  8 426610914    136570… 2021-02-27 17:03:23 Furlow… "@IB… Twitt…     280 136568…
##  9 426610914    136567… 2021-02-27 14:41:01 Furlow… "@IB… Twitt…      35 136546…
## 10 1449382200   136566… 2021-02-27 14:17:34 TdiShe… "I a… Twitt…     234 <NA>   
## # … with 392 more rows, 83 more variables: reply_to_user_id <chr>,
## #   reply_to_screen_name <chr>, is_quote <lgl>, is_retweet <lgl>,
## #   favorite_count <dbl>, retweet_count <dbl>, quote_count <lgl>,
## #   reply_count <lgl>, hashtags <lgl>, symbols <lgl>, urls_url <lgl>,
## #   urls_t.co <lgl>, urls_expanded_url <lgl>, media_url <lgl>,
## #   media_t.co <lgl>, media_expanded_url <lgl>, media_type <lgl>,
## #   ext_media_url <lgl>, ext_media_t.co <lgl>, ext_media_expanded_url <lgl>, …

Creating Edgelist

ties_1 <- ngss_tweets |>
  relocate(sender = screen_name,
           target = reply_to_screen_name) |>
  select(sender,
         target,
         created_at,
         text) 

ties_1

## # A tibble: 402 × 4
##    sender       target      created_at          text                            
##    <chr>        <chr>       <dttm>              <chr>                           
##  1 loyr2662     <NA>        2021-02-27 17:33:27 "Switching gears for a bit for …
##  2 loyr2662     <NA>        2021-02-20 20:02:37 "Was just introduced to the Eng…
##  3 Ngss_18      <NA>        2021-02-27 17:04:18 "Rasanya pengen banget punya te…
##  4 Ngss_18      <NA>        2021-02-26 13:51:35 "Kangen mantan boleh gak sih\U0…
##  5 Ngss_18      <NA>        2021-02-27 13:44:20 "Sebenarnya udah pengen ngejauh…
##  6 Ngss_18      <NA>        2021-02-26 13:45:00 "Bingung mau cerita kesiapa;) t…
##  7 Ngss_18      <NA>        2021-02-27 02:52:34 "Bukannya gak boleh liat keatas…
##  8 Furlow_teach IBchemmilam 2021-02-27 17:03:23 "@IBchemmilam @chemmastercorey …
##  9 Furlow_teach IBchemmilam 2021-02-27 14:41:01 "@IBchemmilam @chemmastercorey …
## 10 TdiShelton   <NA>        2021-02-27 14:17:34 "I am so honored and appreciati…
## # … with 392 more rows

Unnesting Usernames and Dropping Na Values

ties<- ties_1 |>
  unnest_tokens(input = target,
                output = receiver,
                to_lower = FALSE) |>
  relocate(sender, receiver) |>
  drop_na()
  

ties

## # A tibble: 119 × 4
##    sender          receiver        created_at          text                     
##    <chr>           <chr>           <dttm>              <chr>                    
##  1 Furlow_teach    IBchemmilam     2021-02-27 17:03:23 "@IBchemmilam @chemmaste…
##  2 Furlow_teach    IBchemmilam     2021-02-27 14:41:01 "@IBchemmilam @chemmaste…
##  3 jeansciqueen    jeansciqueen    2021-02-27 15:30:46 "Wanda, any book you can…
##  4 haruzoee        JAPANFESS       2021-02-27 14:49:31 "@JAPANFESS manga masih …
##  5 betsylawrence82 KATSorg         2021-02-27 14:10:36 "@KATSorg is seeking pre…
##  6 Joelwry_dev     ngss_official   2021-02-27 14:07:06 "@ngss_official Loved d …
##  7 richbacolor     richbacolor     2021-02-27 12:50:58 "The #ngss vans are gonn…
##  8 masamasa_ngss   EoMTcwAOjlyfS05 2021-02-23 16:30:22 "@EoMTcwAOjlyfS05 肝炎の…
##  9 masamasa_ngss   uy_amm          2021-02-21 23:15:50 "@uy_amm 国印？"         
## 10 masamasa_ngss   RuiRui19995101  2021-02-27 05:54:21 "@RuiRui19995101 やろ"   
## # … with 109 more rows

write_csv(ties,"ngss-edgelist.csv")

Creating Node List

actors_1 <- ties |>
  select(sender, receiver) |>
  pivot_longer(cols = c(sender,receiver))

actors_1

## # A tibble: 238 × 2
##    name     value          
##    <chr>    <chr>          
##  1 sender   Furlow_teach   
##  2 receiver IBchemmilam    
##  3 sender   Furlow_teach   
##  4 receiver IBchemmilam    
##  5 sender   jeansciqueen   
##  6 receiver jeansciqueen   
##  7 sender   haruzoee       
##  8 receiver JAPANFESS      
##  9 sender   betsylawrence82
## 10 receiver KATSorg        
## # … with 228 more rows

actors <- actors_1 |>
  select(value) |>
  rename(actors = value) |> 
  distinct()

actors

## # A tibble: 132 × 1
##    actors         
##    <chr>          
##  1 Furlow_teach   
##  2 IBchemmilam    
##  3 jeansciqueen   
##  4 haruzoee       
##  5 JAPANFESS      
##  6 betsylawrence82
##  7 KATSorg        
##  8 Joelwry_dev    
##  9 ngss_official  
## 10 richbacolor    
## # … with 122 more rows

There are 132 unique actors in the network.

Creating a Network Graph

ngss_network <- tbl_graph(edges = ties, 
                          nodes = actors, 
                          directed = TRUE)
ngss_network

## # A tbl_graph: 132 nodes and 119 edges
## #
## # A directed multigraph with 60 components
## #
## # Node Data: 132 × 1 (active)
##   actors         
##   <chr>          
## 1 Furlow_teach   
## 2 IBchemmilam    
## 3 jeansciqueen   
## 4 haruzoee       
## 5 JAPANFESS      
## 6 betsylawrence82
## # … with 126 more rows
## #
## # Edge Data: 119 × 4
##    from    to created_at          text                                          
##   <int> <int> <dttm>              <chr>                                         
## 1     1     2 2021-02-27 17:03:23 @IBchemmilam @chemmastercorey I’m familiar w/…
## 2     1     2 2021-02-27 14:41:01 @IBchemmilam @chemmastercorey How well does t…
## 3     3     3 2021-02-27 15:30:46 Wanda, any book you can recommend in science/…
## # … with 116 more rows

autograph(ngss_network)

Measuring Centrality Degree

ngss_network <- ngss_network |>
  activate(nodes) |>
  mutate(in_degree = centrality_degree(mode = "in"),
         out_degree = centrality_degree(mode = "out"))
  as_tibble(ngss_network)

## # A tibble: 132 × 3
##    actors          in_degree out_degree
##    <chr>               <dbl>      <dbl>
##  1 Furlow_teach            0          2
##  2 IBchemmilam             2          0
##  3 jeansciqueen            1          1
##  4 haruzoee                0          1
##  5 JAPANFESS               1          0
##  6 betsylawrence82         0          1
##  7 KATSorg                 1          0
##  8 Joelwry_dev             0          1
##  9 ngss_official           3          1
## 10 richbacolor             1          1
## # … with 122 more rows

Based on the centrality degree measures, it seems as though the users with the highest in-degree measures also tend have the among the highest out-degree, making them transcenders. The top 5 transcenders are:

NGSS_tweeps: 6 in-degree, 5 out-degree
jrosenberg6432: 5 in-degree, 5 out-degree
phenomica: 4 in-degree, 4 out-degree
mrjtweeter: 3 in-degree, 4 out-degree
MsMagaryPhysics: 3 in-degree, 3 out-degree

The top transmitters are:

JP_NGSS: 4 out-degree, 0 in-degree
masamas_ngss: 3 out-degree, 1 in-degree
FRC4thDimension: 3 out-degree, 1 in-degree

The top transceivers are:

lovopium: 3 in-degree, 0 out-degree
miss_schilling: 3 in-degree, 0 out-degree
DrDSteele: 3 in-degree, 1 out-degree
Alikara06557110: 3 in-degree, 1 out-degree

What’s interesting, particularly for the transmitters and transceivers, is that the top accounts have “NGSS” in the name, signifying that they may play a very specific role in spreading the word and leading conversation about NGSS.

ngss_network |> 
  as_tibble() |>
  skim()

Data summary
Name	as_tibble(ngss_network)
Number of rows	132
Number of columns	3
_______________________
Column type frequency:
character	1
numeric	2
________________________
Group variables	None

Variable type: character

skim_variable	n_missing	complete_rate	min	max	empty	n_unique	whitespace
actors	0	1	4	15	0	132	0

Variable type: numeric

skim_variable	n_missing	complete_rate	mean	sd	p0	p25	p50	p75	p100	hist
in_degree	0	1	0.9	1.03	0	0	1	1	6	▇▁▁▁▁
out_degree	0	1	0.9	1.00	0	0	1	1	5	▇▁▁▁▁

Each actor had an average of 2.09 alters.

Visualizations

In-Degree Histogram

ngss_network |> 
  as_tibble() |>
  ggplot() +
  geom_histogram(aes(x = in_degree), color = "pink", fill = "pink") +
  labs(title = "Actor In-Degree Measures")

## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

Based on in-degree, no one received more than 6 replies, a little over 60 actors received at least one reply, around 50 recieved no replies at all and just under 10 received between 2 to 3 replies.

Out-Degree Histogram

ngss_network |> 
  as_tibble() |>
  ggplot() +
  geom_histogram(aes(x = out_degree), color = "orange", fill = "orange") +
  labs(title = "Actor Out-Degree Measures")

## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

Based on out-degree, no one replied to more than 5 tweets, a little of 50 actors replied at least once, around 50 didn’t reply at all and just under 10 had sent at least two to 3 replies.

Betweeness and Closeness Measures

ngss_network <-ngss_network |>
  activate(nodes) |>
  mutate(betweenness = centrality_betweenness(), closeness = centrality_closeness())

ngss_network |>
  as_tibble()

## # A tibble: 132 × 5
##    actors          in_degree out_degree betweenness closeness
##    <chr>               <dbl>      <dbl>       <dbl>     <dbl>
##  1 Furlow_teach            0          2           0     1    
##  2 IBchemmilam             2          0           0   NaN    
##  3 jeansciqueen            1          1           0   NaN    
##  4 haruzoee                0          1           0     1    
##  5 JAPANFESS               1          0           0   NaN    
##  6 betsylawrence82         0          1           0     1    
##  7 KATSorg                 1          0           0   NaN    
##  8 Joelwry_dev             0          1           0     0.333
##  9 ngss_official           3          1           3     1    
## 10 richbacolor             1          1           0   NaN    
## # … with 122 more rows

NGSS_tweeps and ms_hyphenated have the highest betweeness scores, at 22 and 6, respectively. The two play the biggest roles in connecting the network. Giving that ‘NGSS’ is in the NGSS_tweeps name, it is clear that this account is doing a lot of bridging between the actors.

In terms of closeness, many of the actors had a closeness between .5 and 1, which signifying very close proximity between each of the actors.

Sociograms

Out-Degree Sociogram

ggraph(ngss_network, layout = "kk") + 
  geom_node_point(aes(color = out_degree, size = 3)) +
  geom_node_text(aes(label = actors, size = out_degree/2),
                 repel=TRUE) +
  geom_edge_arc(arrow = arrow(length = unit(0.5, 'mm')), 
                 end_cap = circle(1, 'mm'),
                 alpha = .2, size = 0.1) +
  scale_color_gradient(low = "lightpink", high = "deeppink") +
  labs(title = "#NGSS Tweets Sociogram Based on Out-Degree") +
  theme_graph()

## Warning in geom_edge_arc(arrow = arrow(length = unit(0.5, "mm")), end_cap =
## circle(1, : Ignoring unknown parameters: `edge_size`

## Warning: Using the `size` aesthetic in this geom was deprecated in ggplot2 3.4.0.
## ℹ Please use `linewidth` in the `default_aes` field and elsewhere instead.

## Warning in grid.Call(C_stringMetric, as.graphicsAnnot(x$label)): font family
## not found in Windows font database

## Warning in grid.Call(C_stringMetric, as.graphicsAnnot(x$label)): font family
## not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family not found in Windows font database

## Warning in grid.Call(C_stringMetric, as.graphicsAnnot(x$label)): font family
## not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family not found in Windows font database

## Warning: ggrepel: 4 unlabeled data points (too many overlaps). Consider
## increasing max.overlaps

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family not found in Windows font database

In-Degree Sociogram

ggraph(ngss_network, layout = "kk") + 
  geom_node_point(aes(color = in_degree, size = 3)) +
  geom_node_text(aes(label = actors, size = in_degree/2),
                 repel=TRUE) +
  geom_edge_arc(arrow = arrow(length = unit(0.5, 'mm')), 
                 end_cap = circle(1, 'mm'),
                 alpha = .2, size = 0.1) +
  scale_color_gradient(low = "lightblue", high = "darkblue") +
  labs(title = "#NGSS Tweets Sociogram Based on In-Degree") +
  theme_graph()

## Warning in geom_edge_arc(arrow = arrow(length = unit(0.5, "mm")), end_cap =
## circle(1, : Ignoring unknown parameters: `edge_size`

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family not found in Windows font database

## Warning: ggrepel: 3 unlabeled data points (too many overlaps). Consider
## increasing max.overlaps

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family not found in Windows font database

## Warning in grid.Call(C_textBounds, as.graphicsAnnot(x$label), x$x, x$y, : font
## family not found in Windows font database

Key Findings, Final Implications and Future Studies

In this study, the following research questions were addressed:
- Do the actors with high in-degree centrality also tend to have high out-degree centrality?
- Which actors serve as transmitters, transceivers and transcenders in the NGSS tweets network?
Overall, it was found that for the most part, the actors with the highest in-degree centrality measures also had high-out degree measures, making them transcenders. Interestingly enough, the biggest transcender was an account that had “NGSS” in the name, signifying that it is an account specifically designated to interact within the #NGSSTweets space. They were able to start conversation, as well as continue conversation through replying to others. The same was true for the transmitters and transceivers, who’s top accounts also had “NGSS” in the name.
Based on other measures such as closeness and betweeness, and calculations for the average number of alters, it appears that the #NGSSTweets network has actors that are relatively close in proximity, and has a central actor with a large betweeness of 22. On average, each actor is connected to at least 2 alters. Knowing this, this might imply that this particular Twitter network has the potential to spread information to one another with relative ease. With nearly all of the actors having between measures ranging from .5-1.00, this was interesting and may signify that productive conversation and spreading the word and opinions about NGSS is happening and possible, though many of the actors seem to either be affiliated with NGSS in some way or are teachers that are already familiar with the standards. This could potentially be an effective platform for gaining more information and opinions regarding the NGSS for researchers and those interested in joining the conversation surrounding the standards.
In a future study, I am interested in comparing the #NGSSTweets network to that of the #CCSSTweets network to explore more about the cliques and subgroups and other centrality measures that exist and why they are organized the way they are. Both of these state standards have been portrayed differently in the media and government, and I am really interested to see how that has had an impact on the networks’ individual structures.