Skip to contents

Create KDE, CDF, or empirical CDF plots

Source: R/density_plot.R
density_plot.Rd

Creates a kernel density estimate, empirical cumulative distribution function, or cumulative distribution function plot of selected variable. Grouping, filtering, and several plot options are available.

Usage

density_plot(
  dat,
  project,
  var,
  type = "kde",
  group = NULL,
  combine = TRUE,
  date = NULL,
  filter_date = NULL,
  date_value = NULL,
  filter_by = NULL,
  filter_value = NULL,
  filter_expr = NULL,
  facet_by = NULL,
  conv = "none",
  tran = "identity",
  format_lab = "decimal",
  scale = "fixed",
  bw = 1,
  position = "identity",
  pages = "single"
)

Arguments

dat

Primary data containing information on hauls or trips. Table in the FishSET database contains the string 'MainDataTable'.

project

String, name of project.

var

String, name of variable to plot.

type

String, type of density plot. Options include "kde" (kernel density estimate), "ecdf" (empirical cdf), "cdf" (cumulative distribution function), or "all" (all plot types). Two or more plot types can be chosen.

group

Optional, string names of variables to group by. If two or grouping variables are included, the default for "cdf" and "ecdf" plots is to not combine groups. This can be changed using combine = TRUE. "kde" plots always combine two or more groups. "cdf" and "ecdf" plots can use up to two grouping variables if combine = FALSE: the first variable is represented by color and second by line type.

combine

Logical, whether to combine the variables listed in group for plot.

date

Date variable from dat used to subset and/or facet the plot by.

filter_date

The type of filter to apply to `MainDataTable`. To filter by a range of dates, use filter_date = "date_range". To filter by a given period, use "year-day", "year-week", "year-month", "year", "month", "week", or "day". The argument date_value must be provided.

date_value

This argument is paired with filter_date. To filter by date range, set filter_date = "date_range" and enter a start- and end-date into date_value as a string: date_value = c("2011-01-01", "2011-03-15").

To filter by period (e.g. "year", "year-month"), use integers (4 digits if year, 1-2 digits if referencing a day, month, or week). Use a vector if filtering by a single period: date_filter = "month" and date_value = c(1, 3, 5). This would filter the data to January, March, and May.

Use a list if using a year-period type filter, e.g. "year-week", with the format: list(year, period). For example, filter_date = "year-month" and date_value = list(2011:2013, 5:7) will filter the data table from May through July for years 2011-2013.

filter_by

String, variable name to filter `MainDataTable` by. the argument filter_value must be provided.

filter_value

A vector of values to filter `MainDataTable` by using the variable in filter_by. For example, if filter_by = "GEAR_TYPE", filter_value = 1 will include only observations with a gear type of 1.

filter_expr

String, a valid R expression to filter `MainDataTable` by.

facet_by

Variable name to facet by. This can be a variable that exists in dat or a variable created by density_plot() such as "year", "month", or "week". date is required if facetting by period.

conv

Convert catch variable to "tons", "metric_tons", or by using a function entered as a string. Defaults to "none" for no conversion.

tran

String; name of function to transform variable, for example "log" or "sqrt".

format_lab

Formatting option for x-axis labels. Options include "decimal" or "scientific".

scale

Scale argument passed to facet_grid. Defaults to "fixed". Other options include "free_y", "free_x", and "free".

bw

Adjusts KDE bandwidth. Defaults to 1.

position

The position of the grouped variable for KDE plot. Options include "identity", "stack", and "fill".

pages

Whether to output plots on a single page ("single", the default) or multiple pages ("multi").

Value

denstiy_plot() can return up to three plots in a single call. When pages = "single" all plots are combined and stacked vertically.

pages = "multi" will return separate plots.

Details

The data can be filtered by date or by variable (see filter_date and filter_by). If type contains "kde" or "all" then grouping variables are automatically combined. Any variable in dat can be used for faceting, but "year", "month", or "week" are also available if date is provided.

Examples

if (FALSE) {

density_plot(pollockMainDataTable, "pollock", var = "OFFICIAL_TOTAL_CATCH_MT",
             type = c("kde", "ecdf"))

# facet 
density_plot(pollockMainDataTable, "pollock", var = "OFFICIAL_TOTAL_CATCH_MT",
             type = c("kde", "ecdf"), facet_by = "GEAR_TYPE")

# filter by period
density_plot(pollockMainDataTable, "pollock", var = "OFFICIAL_TOTAL_CATCH_MT", 
             type = "kde", date = "FISHING_START_DATE", filter_date = "year-month", 
             filter_value = list(2011, 9:11))
}