Two_mode_network_viz
Tiangeng Lu
2023-08-08
1990s
df_two_mode_90 <- read.csv('two_mode_90.csv', row.names = 1)
df_two_mode_90 <- df_two_mode_90[,!colnames(df_two_mode_90) %in% "pol"]
print(paste("There are", nrow(df_two_mode_90), "items."))## [1] "There are 133 items."#### summary statistics ####
mean(rowSums(df_two_mode_90))## [1] 1.684211paste("On average, an article contains", round(mean(rowSums(df_two_mode_90)),2), "keyword themes")## [1] "On average, an article contains 1.68 keyword themes"# share of articles w/o themes, calculated from row sum
paste(round(table(rowSums(df_two_mode_90))[1] / nrow(df_two_mode_90),4), "articles do not contain any keyword themes.")## [1] "0.0977 articles do not contain any keyword themes."df_two_mode_90_simple <- df_two_mode_90[rowSums(df_two_mode_90) > 0,]
print(paste(nrow(df_two_mode_90_simple), "are kept for network visualization."))## [1] "120 are kept for network visualization."# How many rows in difference?
print(nrow(df_two_mode_90) - nrow(df_two_mode_90_simple))## [1] 13two_mode_90_graph <- graph_from_incidence_matrix(as.matrix(df_two_mode_90_simple))
# How many nodes?
V_count <- length(V(two_mode_90_graph))Validation
# Confirms that column names (features) are among the node names
colnames(df_two_mode_90) %in% V(two_mode_90_graph)$name## [1] TRUE TRUE TRUE TRUE TRUErownames(df_two_mode_90)[!rownames(df_two_mode_90) %in% V(two_mode_90_graph)$name]## [1] "91_1" "92_2" "95_6" "96_5" "96_10" "97_9" "99_12" "00_4" "00_5"
## [10] "00_11" "00_14" "00_18" "00_29"# V()$type returns FALSE or TRUE, FALSE are the row names
table(V(two_mode_90_graph)$type)##
## FALSE TRUE
## 120 5V(two_mode_90_graph)$label <- ""
# show feature labels
V(two_mode_90_graph)[V(two_mode_90_graph)$type == T]$label <- V(two_mode_90_graph)[V(two_mode_90_graph)$type == T]$name
#### COLOR ####
V(two_mode_90_graph)$color <- c("orange","lightsteelblue")[V(two_mode_90_graph)$type + 1]
#### DEGREE (could impact SIZE) ####
feature_degrees_two_mode_90_graph <- degree(two_mode_90_graph, v = V(two_mode_90_graph)[V(two_mode_90_graph)$type])
log_feature_degrees_two_mode_90_graph <- log(feature_degrees_two_mode_90_graph + 0.1)
feature_degree_range <- max(log_feature_degrees_two_mode_90_graph) - min(log_feature_degrees_two_mode_90_graph)
print(paste("The range of log-transformed degree of features is:", round(feature_degree_range,2)))## [1] "The range of log-transformed degree of features is: 1.63"#### SIZE ####
## set initial sizes for ids and features
V(two_mode_90_graph)$size <- c(2, 20)[V(two_mode_90_graph)$type + 1]
## overwrite node size of features, depending on their degree
V(two_mode_90_graph)$size[V(two_mode_90_graph)$type] <- 15 + log_feature_degrees_two_mode_90_graph
# check feature size
V(two_mode_90_graph)$size[V(two_mode_90_graph)$type]## [1] 18.04927 19.63570 18.74005 18.64021 18.00072Visualization
par(mar = c(1, 1, 3, 1))
set.seed(1234)
plot(two_mode_90_graph, vertex.frame.color = NA, vertex.label.cex = 1, vertex.label.font = 2, layout = layout_nicely)
2000s
df_two_mode_00 <- read.csv('two_mode_00.csv', row.names = 1)
df_two_mode_00 <- df_two_mode_00[,!colnames(df_two_mode_00) %in% "pol"]
print(paste("There are", nrow(df_two_mode_00), "items."))## [1] "There are 494 items."#### summary statistics ####
mean(rowSums(df_two_mode_00))## [1] 1.65587paste("On average, an article contains", round(mean(rowSums(df_two_mode_00)),2), "keyword themes")## [1] "On average, an article contains 1.66 keyword themes"# share of articles w/o themes, calculated from row sum
paste(round(table(rowSums(df_two_mode_00))[1] / nrow(df_two_mode_00),4), "articles do not contain any keyword themes.")## [1] "0.1215 articles do not contain any keyword themes."df_two_mode_00_simple <- df_two_mode_00[rowSums(df_two_mode_00) > 0,]
print(paste(nrow(df_two_mode_00_simple), "are kept for network visualization."))## [1] "434 are kept for network visualization."# How many rows in difference?
print(nrow(df_two_mode_00) - nrow(df_two_mode_00_simple))## [1] 60two_mode_00_graph <- graph_from_incidence_matrix(as.matrix(df_two_mode_00_simple))
# How many nodes?
V_count <- length(V(two_mode_00_graph))
# Confirms that column names (features) are among the node names
colnames(df_two_mode_00) %in% V(two_mode_00_graph)$name## [1] TRUE TRUE TRUE TRUE TRUE# Confirms that row names (ids) are among the node names
rownames(df_two_mode_00)[!rownames(df_two_mode_00) %in% V(two_mode_00_graph)$name]## [1] "01_7" "01_16" "01_22" "02_4" "02_12" "02_14" "02_16" "03_12"
## [9] "03_14" "04_10" "05_1" "05_2" "05_3" "05_22" "05_26" "06_1"
## [17] "06_2" "06_5" "06_15" "06_21" "06_23" "06_36" "06_37" "07_25"
## [25] "07_35" "07_51" "08_18" "08_22" "08_27" "08_32" "08_33" "08_40"
## [33] "08_51" "08_54" "08_61" "09_4" "09_9" "09_13" "09_45" "09_55"
## [41] "09_59" "10_6" "10_9" "10_10" "10_15" "10_21" "10_23" "10_33"
## [49] "10_35" "10_36" "10_48" "10_61" "10_63" "10_64" "10_68" "10_69"
## [57] "10_70" "10_93" "10_100" "10_103"length(rownames(df_two_mode_00) %in% V(two_mode_00_graph)$name)## [1] 494# V()$type returns FALSE or TRUE, FALSE are the row names
table(V(two_mode_00_graph)$type)##
## FALSE TRUE
## 434 5#### LABELS ####
# default no node labels
V(two_mode_00_graph)$label <- ""
# show feature labels
V(two_mode_00_graph)[V(two_mode_00_graph)$type == T]$label <- V(two_mode_00_graph)[V(two_mode_00_graph)$type == T]$name
#### COLOR ####
V(two_mode_00_graph)$color <- c("orange","lightsteelblue")[V(two_mode_00_graph)$type + 1]
#### DEGREE (could impact SIZE) ####
feature_degrees_two_mode_00_graph <- degree(two_mode_00_graph, v = V(two_mode_00_graph)[V(two_mode_00_graph)$type])
log_feature_degrees_two_mode_00_graph <- log(feature_degrees_two_mode_00_graph + 0.1)
feature_degree_range <- max(log_feature_degrees_two_mode_00_graph) - min(log_feature_degrees_two_mode_00_graph)
print(paste("The range of log-transformed degree of features is:", round(feature_degree_range,2)))## [1] "The range of log-transformed degree of features is: 2.08"#### SIZE ####
## set initial sizes for ids and features
V(two_mode_00_graph)$size <- c(2, 20)[V(two_mode_00_graph)$type + 1]
## overwrite node size of features, depending on their degree
V(two_mode_00_graph)$size[V(two_mode_00_graph)$type] <- 15 + log_feature_degrees_two_mode_00_graph
# check feature size
V(two_mode_00_graph)$size[V(two_mode_00_graph)$type]## [1] 19.83708 20.79332 19.99111 20.17105 18.71601par(mar = c(1, 1, 3, 1))
set.seed(1234)
#### Visualize ####
plot(two_mode_00_graph, vertex.frame.color = NA, vertex.label.cex = 1, vertex.label.font = 2, layout = layout_nicely)
2010s
df_two_mode_10 <- read.csv('two_mode_10.csv', row.names = 1)
df_two_mode_10 <- df_two_mode_10[,!colnames(df_two_mode_10) %in% "pol"]
print(paste("There are", nrow(df_two_mode_10), "items."))## [1] "There are 1519 items."#### summary statistics ####
mean(rowSums(df_two_mode_10))## [1] 1.545095paste("On average, an article contains", round(mean(rowSums(df_two_mode_10)),2), "keyword themes")## [1] "On average, an article contains 1.55 keyword themes"# share of articles w/o themes, calculated from row sum
paste(round(table(rowSums(df_two_mode_10))[1] / nrow(df_two_mode_10),4), "articles do not contain any keyword themes.")## [1] "0.1422 articles do not contain any keyword themes."df_two_mode_10_simple <- df_two_mode_10[rowSums(df_two_mode_10) > 0,]
print(paste(nrow(df_two_mode_10_simple), "are kept for network visualization."))## [1] "1303 are kept for network visualization."# How many rows in difference?
print(nrow(df_two_mode_10) - nrow(df_two_mode_10_simple))## [1] 216two_mode_10_graph <- graph_from_incidence_matrix(as.matrix(df_two_mode_10_simple))
# How many nodes?
V_count <- length(V(two_mode_10_graph))
# Confirms that column names (features) are among the node names
colnames(df_two_mode_10) %in% V(two_mode_10_graph)$name## [1] TRUE TRUE TRUE TRUE TRUE# Confirms that row names (ids) are among the node names, how many ids are not associated with any features?
length(rownames(df_two_mode_10)[!rownames(df_two_mode_10) %in% V(two_mode_10_graph)$name])## [1] 216# V()$type returns FALSE or TRUE, FALSE are the row names
table(V(two_mode_10_graph)$type)##
## FALSE TRUE
## 1303 5#### LABELS ####
# default no node labels
V(two_mode_10_graph)$label <- ""
# show feature labels
V(two_mode_10_graph)[V(two_mode_10_graph)$type == T]$label <- V(two_mode_10_graph)[V(two_mode_10_graph)$type == T]$name
#### COLOR ####
V(two_mode_10_graph)$color <- c("orange","lightsteelblue")[V(two_mode_10_graph)$type + 1]
#### DEGREE (could impact SIZE) ####
feature_degrees_two_mode_10_graph <- degree(two_mode_10_graph, v = V(two_mode_10_graph)[V(two_mode_10_graph)$type])
log_feature_degrees_two_mode_10_graph <- log(feature_degrees_two_mode_10_graph + 0.1)
feature_degree_range <- max(log_feature_degrees_two_mode_10_graph) - min(log_feature_degrees_two_mode_10_graph)
print(paste("The range of log-transformed degree of features is:", round(feature_degree_range,2)))## [1] "The range of log-transformed degree of features is: 2.02"#### SIZE ####
## set initial sizes for ids and features
V(two_mode_10_graph)$size <- c(2, 20)[V(two_mode_10_graph)$type + 1]
## overwrite node size of features, depending on their degree
V(two_mode_10_graph)$size[V(two_mode_10_graph)$type] <- 15 + log_feature_degrees_two_mode_10_graph
# check feature size
V(two_mode_10_graph)$size[V(two_mode_10_graph)$type]## [1] 21.05467 21.84599 20.93251 21.17191 19.82911par(mar = c(1, 1, 3, 1))
set.seed(1234)
plot(two_mode_10_graph, vertex.frame.color = NA, vertex.label.cex = 1.8, vertex.label.font = 2, layout = layout_with_lgl)