Merge pull request #14 from mattansb/2021

2021

01 path analysis/Appendix - advanced plotting.R

@@ -0,0 +1,154 @@
+library(lavaan)
+library(tidySEM)
+library(dplyr)
+
+income_psych <- read.csv("income_psych.csv")
+
+mediation_model <- '
+  neg_mood ~ anxiety
+    income ~ anxiety + neg_mood
+'
+fit <- sem(mediation_model, data = income_psych)
+
+## See below for how "plot1.png" and "plot2.png" were made.
+
+
+# Changing Layout ---------------------------------------------------------
+
+lay <- get_layout(
+  NA,        "neg_mood", NA,
+  "anxiety", NA,         "income",
+  rows = 2
+)
+
+graph_sem(fit, 
+          layout = lay, angle = 90)
+
+
+
+# Changing Nodes (Vars) ---------------------------------------------------
+
+# Instead of directly plotting, we will use `prepare_graph()` as an intermediary
+# step.
+
+g <- prepare_graph(fit, 
+                   layout = lay, angle = 90)
+
+# we can now see and change node data:
+nodes(g)
+# as you can see, there are many things you can change here!
+# You can also add color, fill, size, linetype and alpha.
+
+
+nodes(g) <- nodes(g) %>% 
+  mutate(label = c("Anxiety", "Income", "Negative\nMood")) # \n is a break line
+
+plot(g)
+
+
+# Changing Edges (Paths) --------------------------------------------------
+
+
+g <- prepare_graph(fit, 
+                   digits = 3, # if we want to change the digits
+                   layout = lay, angle = 90)
+
+
+# And as we did with the nodes, we can see and change edge data:
+edges(g)
+
+
+# as you can see, there are many things you can change here!
+# You can also add color, size, linetype and alpha.
+
+
+edges(g) <- edges(g) %>% 
+  mutate(
+    # set label:
+    label = paste0(est_std, "\n", confint_std),
+    # set color:
+    color = case_when(
+      to == from ~ "black",
+      as.numeric(est_std) > 0 ~ "green",
+      as.numeric(est_std) < 0 ~ "red",
+      TRUE ~ "black"
+    ),
+    # set linetype
+    linetype = case_when(
+      as.numeric(pval_std) > 0.05 ~ "dashed",
+      TRUE ~ "solid"
+    )
+  )
+
+# (This will all be easier in future versions of the package.)
+
+
+plot(g)
+
+# There are many more options, you can read more here:
+# https://cjvanlissa.github.io/tidySEM/articles/Plotting_graphs.html
+
+
+# Making "plot1" ----------------------------------------------------------
+
+
+mediation_model <- '
+  neg_mood ~ anxiety
+    income ~ anxiety + neg_mood
+'
+
+fit <- sem(mediation_model, data = income_psych)
+
+lay <- get_layout(
+  NA,        "neg_mood", NA,
+  "anxiety", NA,         "income",
+  rows = 2
+)
+
+
+g1 <- prepare_graph(fit, layout = lay, angle = 90)
+
+nodes(g1) <- nodes(g1) %>% 
+  mutate(label = c("Anxiety", "Income", "Negative\nMood")) # \n is a break line
+
+edges(g1) <- edges(g1) %>% 
+  filter(to != from) %>% 
+  mutate(label = "")
+
+
+plot(g1)
+ggplot2::ggsave("plot1.png", height = 4, width = 6)
+
+
+# Making "plot2" ----------------------------------------------------------
+
+
+mediation_model <- '
+  neg_mood ~ anxiety
+   shyness ~ anxiety
+    income ~ anxiety + neg_mood + shyness
+'
+
+fit <- sem(mediation_model, data = income_psych)
+
+
+lay <- get_layout(
+  NA,        "neg_mood", NA,
+  "anxiety", NA,         "income",
+  NA,        "shyness",  NA,
+  rows = 3
+)
+
+
+g2 <- prepare_graph(fit, layout = lay, angle = 90)
+
+
+nodes(g2) <- nodes(g2) %>% 
+  mutate(label = c("Anxiety", "Income", "Negative\nMood", "Shyness")) # \n is a break line
+
+edges(g2) <- edges(g2) %>% 
+  filter(to != from) %>% 
+  mutate(label = "")
+
+plot(g2)
+ggplot2::ggsave("plot2.png", height = 4, width = 6)