LightLogR (version 0.9.2)

summarize_numeric: Summarize numeric columns in dataframes to means

Description

This simple helper function was created to summarize episodes of gaps, clusters, or states, focusing on numeric variables. It calculates mean values for all numeric columns and handles Duration objects appropriately.

Despite its name, the function actually summarizes all double columns, which is more inclusive compared to just numeric columns.

Usage

summarize_numeric(
  data,
  remove = NULL,
  prefix = "mean_",
  na.rm = TRUE,
  complete.groups.on = NULL,
  add.total.duration = TRUE,
  durations.dec = 0,
  Datetime2Time = TRUE
)
summarise_numeric(
  data,
  remove = NULL,
  prefix = "mean_",
  na.rm = TRUE,
  complete.groups.on = NULL,
  add.total.duration = TRUE,
  durations.dec = 0,
  Datetime2Time = TRUE
)

Value

A dataframe containing the summarized metrics.

Arguments

data

A dataframe containing numeric data, typically from extract_clusters() or extract_gaps().

remove

Character vector of columns removed from the summary.

prefix

A prefix to add to the column names of summarized metrics. Defaults to "mean_".

na.rm

Whether to remove NA values when calculating means. Defaults to TRUE.

complete.groups.on

Column name that, together with grouping variables, can be used to provide a complete set. For example, with extract_clusters(), some days might not have clusters. They do not show up in the summary output then. If it is important however, to consider that there are zero instances, one could extract the complete set of clusters and non-clusters, and then set is.cluster in this argument, which would then show zero clusters for those days.

add.total.duration

Logical, whether the total duration for a given group should be calculated. Only relevant if a column duration is part of the input data.

durations.dec

Numeric of number of decimals for the mean calculation of durations and times. Defaults to 0.

Datetime2Time

Logical of whether POSIXct columns should be transformed into hms(time) columns, which is usually sensible for averaging (default is TRUE). Calls Datetime2Time() with default settings (all POSIXct are affected).

Examples

# Extract clusters and summarize them
dataset <-
sample.data.environment %>%
aggregate_Datetime(unit = "15 mins") |>
extract_clusters(MEDI > 1000)

#input to summarize_numeric
dataset |> utils::head()
#output of summarize_numeric (removing state.count and epoch from the summary)
dataset |> summarize_numeric(c("state.count", "epoch"))