I’m building a small shiny app to demo some causal inference concepts and I want to split all the independent variables from a linear regression formula and show them in multiple plots.

Strangely, when I look through the list of independent variables, shiny renders the last of them twice rather than rendering the first and then the second.

Here’s the UI part of my code:

  ui <- fluidPage(
    textInput("regression", "Regression formula", ""),
    verbatimTextOutput("summary"),
    
    plotOutput("plot1", width=fig.width, height=fig.height),
    plotOutput("plot2", width=fig.width, height=fig.height),
    plotOutput("plot3", width=fig.width, height=fig.height),
  )

and here’s where I split the formula into multiple parts and try to render all the independent variables:

      first_split <- unlist(strsplit(formula, "[~]"))
      dependent <- str_trim(first_split[1])
      independent <- unlist(lapply(strsplit(first_split[2], "[+]"), str_trim))
      
      for( i in 1:length(independent))
      {
        whichplot <- paste("plot", toString(i), sep="")
        
        print(whichplot) # this prints "plot1" and "plot2"
        print(independent[i]) # this prints "prek" and "new_books"
        
        output[[whichplot]] <- renderPlot({
          ggplot(reading_scores, aes(x = .data[[dependent]])) +
            geom_histogram(aes(color = as.factor(.data[[independent[i]]]), fill = as.factor(.data[[independent[i]]])), 
                           position = "identity", bins = 30, alpha = 0.2) +
            scale_color_manual(values = c("#00AFBB", "#E7B800")) +
            scale_fill_manual(values = c("#00AFBB", "#E7B800"))
        })
      }

When I pass “scores ~ prek + new_books” as the formula, what should happen is that I get a hist of both ‘prek’ and ‘new_books’ but instead I just get ‘new_books’ twice. I suspect that shiny is re-rendering the plot but I can’t figure out why.