plot_word_network: Plot Word Network
In nlanderson9/languagePredictR: Predict Outcomes from Natural Language

Description Usage Arguments Value Examples

Plots a word network of adjacent words.

plot_word_network(
  input,
  model = NULL,
  topX = 100,
  graphIndividual = TRUE,
  graphCombined = FALSE,
  directed = FALSE,
  removeVerticesBelowDegree = 2,
  clusterType = "none",
  clusterNodeMethod = "infomap",
  plotUnclusteredNetwork = TRUE,
  plotClusteredNetwork = TRUE,
  plotIndividualClusters = TRUE,
  plotIndividualClusterFacet = TRUE,
  plotClusterLegend = TRUE,
  plotNodeColors = TRUE,
  edgeColor = "darkgray",
  edgeAlpha = 0.5,
  edgeCurve = 0.15,
  modelNodeColors = c("lightblue", "orange"),
  modelNodeSizeRange = c(5, 10),
  nodeLabelSize = 1,
  nodeLabelColor = "black",
  plotTitle = NULL,
  layout = NULL
)

`input`	An input dataframe, typically the output from the `node_edge` function
`model`	Optional - if `node_edge` used a model as input, the same model can be provided here for extra functionality
`topX`	The number of word pairs to include in the graphed network. Chosen word pairs are selected from those with the greatest number of co-occurrences. Defaults to 100. If NULL, all pairs will be used.
`graphIndividual`	If TRUE, individual graphs are produced for both outcomes of the model. Default is TRUE.
`graphCombined`	If TRUE, a network is graphed based on the entire language corpus. Default is FALSE.
`directed`	Determines if the network is directed (direction of edges matters) or not. Defaults to FALSE (the output from `node_edge` does not yield directional edge information, so only change this if using your own dataframe).
`removeVerticesBelowDegree`	An integer which determines the minimum number of edges a node must have to be included. Default is 2.
`clusterType`	The type of clustering to perform. "node" clusters by nodes (using the method defined by `clusterNodeMethod`), "edge" clusters by edges (using the `lincomm` package. Defaults to "none".
`clusterNodeMethod`	If clustering by "node", this determines the method used. Options are the same clustering options given in the `igraph` package.
`plotUnclusteredNetwork`	If TRUE, the network is plotted with no clustering displayed. Defaults to TRUE.
`plotClusteredNetwork`	If TRUE, the network is plotted with clustering displayed (shaded regions for "node" clustering, colored edges for "edge" clustering). Defaults to TRUE.
`plotIndividualClusters`	If TRUE, each cluster is plotted in a separate graph. Defaults to TRUE.
`plotIndividualClusterFacet`	If TRUE, each cluster is plotted in a faceted section of a single graph. Defaults to TRUE.
`plotClusterLegend`	If TRUE, plots a legend with cluster numbers and corresponding colors. Defaults to TRUE.
`plotNodeColors`	If TRUE, nodes are colored in based on the model provided.
`edgeColor`	The color of the edges. Default is "darkgray".
`edgeAlpha`	The alpha of the edges. Default is 0.5.
`edgeCurve`	If greater than 0, edges will be curved with a radius corresponding to the value. Default is 0.15. A value of 0 yields straight edges.
`modelNodeColors`	The color shading for nodes that are predictive words in the provided model. Must be a vector of two values. Defaults to c("lightblue", "orange").
`modelNodeSizeRange`	The sizing for nodes that are predictive words in the provided model. Must be a vector of two values (the minimum plotted size and maximum plotted size). Defaults to c(5, 10).
`nodeLabelSize`	The size of the text for node labels. Defaults to 1.
`nodeLabelColor`	The color of the node labels. Defaults to "black".
`plotTitle`	The title of the plot(s). If a model is used, it must be a vector of three strings. If not, it must be a single string.
`layout`	An `igraph` layout object, helpful if you want to re-graph the same network multiple times with the same layout

A list containing the igraph network object, igraph layout, and clustering data (if applicable) - this can be used with the plot_cluster function, or graphed manually with the igraph package

## Not run: 
movie_review_data1$cleanText = clean_text(movie_review_data1$text)

# Using language to predict "Positive" vs. "Negative" reviews
movie_model_valence = language_model(movie_review_data1,
                                     outcomeVariableColumnName = "valence",
                                     outcomeVariableType = "binary",
                                     textColumnName = "cleanText")

node_edge_table = node_edge(movie_model_valence)
plot_word_network(node_edge_table)

## End(Not run)