## in server.r
  output$null_plot <- renderDygraph({
    ## don't output anything unless you have the data ready
    validate(
      need(casualImpactData(), "Model Working")
    )    
    
    ## the data for the plot is in here
    ci <- casualImpactData()$series
    
    ## get the start and end from the user input in ui.r
    start <- input$range_date[1]
    end   <- input$range_date[2]
    
    ## we need to convert the dataframe into a timeseries for dygraph
    orderme <- seq(as.Date(start), as.Date(end), by=1)
    ci <- xts(ci, orderme)
    
    ## the dygraph output
    dygraph(data=ci[,c('response', 'point.pred', 'point.pred.lower', 'point.pred.upper')], 
            main="Expected (95% confidence level) vs Observed", group="ci") %>%
      dyRangeSelector(dateWindow = c(input$event_date - 7, input$range_date[2])) %>%
      dyEvent(date = input$event_date, "Event") %>%
      dySeries(c('point.pred.lower', 'point.pred','point.pred.upper'), label='Expected') %>%
      dySeries('response', label="Observed")
  })