plot_outliers: Plot outliers
In scoutiii/HTSoutliers: Hrydological and Time Series Outlier Detection Methods

Description Usage Arguments Details Value Examples

Once you have flagged outliers using any model (randomForest for example), you can plot the outliers that have already been flagged vs the one you have flagged

1	plot_outliers(data, id)

`data`	full data frame with at least SNWD, OUTLIER_FINAL, ID, DATE, OUTLIER_PRED
`id`	the id for the station you want to plot

This will show known outliers as triangles, and points you predict to be outliers as blue. So a blue triangle is a point you marked as an outlier that we also found to be an outlier. A green triangle is a point we found to be and outlier and you did not. A blue circle is a point you flagged as an outlier that we did not. And a green circle is a point neither you nor we flagged.

Note that you must have at least SNWD, OUTLIER_FINAL, ID, DATE, and OUTLIER_PRED in order to use this function. You will need to manually create the OUTLIER_PRED variable

plotly plot

## Not run: 
  # See the readme for a better example.
  # gets the ids of stations which had predicted outliers
  flagged_ids <- dplyr::filter(test_data, OUTLIER_PRED == 1) %>%
    dplyr::select(ID) %>%
    dplyr::distinct()

  # plots the first 20 stations
  for (i in 1:20) {
    print(flagged_ids[i, "ID"])
    print(plot_outliers(test_data, as.character(flagged_ids[i,"ID"])))
  }

## End(Not run)