Type: Package
Title: Wrangling Longitudinal Survival Data
Version: 1.0.1
Description: Streamlines the process of transitioning between data formats commonly used in survival analysis. Functions convert longitudinal data between formats used as input for survival models as well as support overall preparation. Users are able to focus on model building rather than data wrangling.
URL: https://github.com/ci2131a/wlsd
BugReports: https://github.com/ci2131a/wlsd/issues
License: GPL-3
Encoding: UTF-8
LazyData: true
Depends: R (≥ 3.5.0)
Imports: stats
Suggests: knitr, rmarkdown, testthat (≥ 3.0.0)
VignetteBuilder: knitr
RoxygenNote: 7.3.3
Config/testthat/edition: 3
NeedsCompilation: no
Packaged: 2026-02-01 06:00:10 UTC; charl
Author: Charles Ingulli [aut, cre]
Maintainer: Charles Ingulli <charlesfi@outlook.com>
Repository: CRAN
Date/Publication: 2026-02-04 19:50:02 UTC

wlsd: Wrangling Longitudinal Survival Data

Description

logo

Streamlines the process of transitioning between data formats commonly used in survival analysis. Functions convert longitudinal data between formats used as input for survival models as well as support overall preparation. Users are able to focus on model building rather than data wrangling.

Author(s)

Maintainer: Charles Ingulli charlesfi@outlook.com

See Also

Useful links:

Low Back Pain Data Set

Description

A long format data set from a longitudinal study of low back pain (LBP) on midwestern manufacturing workers.

Usage

LBP

Format

A data frame on the following variables:

Variable Description Class
sid: The subject identification variable for individuals. Factor
Baseline.date: The date of baseline visit or enrollment of individuals into the study. Date
Date: The calendar time of follow-up visit. Date
time_to_row: The number of days between the current follow-up visit and the baseline date. Integer
case.lbp: A status indicator for individuals possessing any LBP (0 for no and 1 for yes). Integer
case.med: A status indicator determining whether indviduals are taking medication for LBP (0 for no and 1 for yes). Integer
case.sc: A status indicator to determine whether individuals are seeking care for LBP (0 for no and 1 for yes). Integer
case.ls: A status indicator to determine whether individuals have lost time from work due to LBP (0 for no and 1 for yes). Integer
gender: The gender of the individual (either M for Male or F for Female). Factor
age: The age of the individual at baseline visit in years. Numeric
weight: The weight of individuals in lbs. Integer
height: The height of individuals in inches. Integer
raceth: A categorical variable to determine the race/ethnicity of individuals (0 = White; 1 = Hispanic/Latino; 2 = Black; 3 = Asian; 4 = Native Hawaiian or Pacific Islander; 5 = Native American or Native Alaskan; 6 = Other/declined). Factor
smoking: A smoking indicator variable (0 = Smoked less than 100 cigarettes in life; 1 = smoked in the past, but no longer, 2 = currently smoke). Factor
comptenure: A categorical variable to determine length of time at the current company (0 = less than 3 months; 1 = 3 months to 1 year; 2 = 1 year to 3 years; 3 = 3 years to 5 years; 4 = 5 years to 10 years; 5 = 10 or more years). Factor
jobtenure: A categorical variable to determine length of time in their current job 0 = less than 3 months; 1 = 3 months to 1 year; 2 = 1 year to 3 years; 3 = 3 years to 5 years; 4 = 5 years to 10 years; 5 = 10 or more years. Factor
control.order: A categorical variable to determine how much control individuals have over the order in which they complete tasks (0 = "Very Much", 1 = "Much", 2="Moderate Amounts", 3="A Little", 4="Very Little"). Factor
control.pace: A categorical variable to determine how much control individuals have over the pace in which they complete tasks (0 = "Very Much", 1 = "Much", 2="Moderate Amounts", 3="A Little", 4="Very Little"). Factor
control.breaks: A categorical variable to determine the amount of control individuals have in taking breaks between completing tasks (0 = "Very Much", 1 = "Much", 2="Moderate Amounts", 3="A Little", 4="Very Little"). Factor
supervisor.support: A categorical variable determining how much support individuals feel they receive from their supervisor (0="Almost Always", 1="Some of the Time", 2="Hardly Ever"). Factor
coworker.support: A categorical variable determining how much support individuals feel they receive from their coworkers (0="Almost Always", 1="Some of the Time", 2="Hardly Ever"). Factor
job.satisfied: A categorical variable to determine whether individuals feel satisfied with their current job (0="Very Satisfied", 1="Somewhat Satisfied", 2="A Little Satisfied", 3="Not at all Satisfied"). Factor
bmi: The calculated body mass index (BMI) of individuals based on height and weight. Numeric

Details

Data set construction was done through the consolidation of various source files pulled from the original database. The final data frame contains follow-up information for selected individuals. The case definitions assessed over time were case.lbp, case.med, case.sc, and case.lt. Column time_to_row is constructed using the Baseline.date and Date columns to calculate the number of days between observations (denoted by rows). All other columns are constant with respect to time. Categorical variables were recorded through self-assessment on the part of the subject. The age and weight variables were able to be physically measured to then be used in calculation of bmi.

Source

LBP Research Consortium, University of Wisconsin-Milwaukee

References

Garg, Arun, Kurt Hegmann, J. Moore, Jay Kapellusch, Matthew Thiese, Sruthi Boda, Parag Bhoyar, Donald Bloswick, Andrew Merryweather, Richard Sesek, Gwen Deckow-Schaefer, James Foster, Eric Wood, Xiaoming Sheng, and Richard Holubkov (2013). Study protocol title: A prospective cohort study of low back pain. BMC Musculoskeletal Disorders 14(84), 84.

Ingulli, Charles. (2020). A Survey of Statistical Methods for Investigating Risk of Low Back Pain in a Cohort of Manufacturing Workers. (85696). [Master's Thesis, American University]

Examples

LBP

Create Baseline Row

Description

Creates a new row of values for subjects representing baseline observations in a data set of follow-up observations.

Usage

basedate(data,id)

Arguments

data

Data frame with relevant columns.

id

Character string of the identification column name in data.

Details

Adds a new row for each level of the id column. Internal functions will try to determine any constant columns by checking for consistency within id groups in order to fill in some of the blanks.

Value

A data frame with added row for each level of id.

Examples

basedate(long_data, "id")

Count Format Data Example

Description

A toy data set in count format.

Usage

count_data

Format

A data frame with 3 rows on the following 5 variables.

id

An identification variable

time

Aggregate time variable

event

Aggregated status indicator variable

var1

First example explanatory variable

var2

Second example explanatory variable

Examples

count_data

Counting Process Format to Long format

Description

Transforms data from counting process format to the long format.

Usage

cp2long(data, id, time1, time2, status = NULL, fill = FALSE)

Arguments

data

A data frame with relevant columns.

id

A character string of the identification variable name in data.

time1

A character string of the first time point variable in data. Represents the left endpoint of the time interval.

time2

A character string of the second time point variable in data. Represents the right endpoint of the time interval.

status

A character string of the status column name in data to be treated as either an event or state.

fill

An optional argument that attempts to fill any NA values in the output for columns that might be constant within id levels.

Details

The data transition consolitdates information from the time1 and time2 argument into a single time column. All other columns are assumed to correspond to the time2 point. Thus, the first row generally consists of NA values. The fill argument will attempt to discern any constant columns within id groups in order to populate that first row.

Value

A data frame in long format.

Examples

cp2long(data = cp_data, id = "id", time1 = "time1", time2 = "time2")

Counting Process Data Example

Description

A toy data set in counting process format.

Usage

cp_data

Format

A data frame with 6 rows on the following 6 variables.

id

An identification variable

time1

Starting time of observation interval

time2

Ending time of observation interval

event

Status indicator variable

var1

First example explanatory variable

var2

Second example explanatory variable

Examples

cp_data

Multiple Event Variables to One State Variable

Description

Converts one or more event columns within a data frame to a single state vector whose values represent combinations of events.

Usage

events2state(data, events, number = TRUE, drop = TRUE, ...)

Arguments

data

A data frame with relevant columns.

events

The names of the event variables as character strings in a vector.

number

A logical argument to determine whether the new state variable should be converted to a number representing the combination of events or left as is. Defaults to TRUE which will convert combinations a numeric. If argument is set to FALSE, the combinations will be left unchanged.

drop

Passed to interaction in order to determine whether unused factors will be excluded from the defining levels. The default is TRUE.

...

Further arguments to be passed to interaction.

Details

For a data frame with the necessary inputs, the function will aggregate values across columns supplied to events through the interaction function. The key for the different combination levels is printed to the console.

Value

Returns the input data frame with an added column called state.

Examples

events2state(data = long_data, events = c("event", "var2"))

Longitudinal to Count format

Description

Aggregates longitudinal data into a count format data set.

Usage

long2count(data, id, event = NULL, state = NULL, FUN, ...)

Arguments

data

A data frame with relevant columns.

id

A character string of the identification variable name in data.

event

The name(s) of the event column(s) in data to be tallied. The name(s) is required to be supplied as a string. The elements of this argument are assumed to be numeric and are summed for each identification level from id.

state

The name of the state variable in data. This argument is used if the event of interest is a numeric or non-numeric series of states. Each of these levels will be tallied for each level of the id.

FUN

The summary function to be applied to all time-depentent columns (wrapper for argument in stats::aggregate). If nothing is supplied, then mean will be used.

...

Additional arguments supplied to stats::aggregate.

Details

The returned data frame aggregates any time-depended values based on row-wise changes within id groups. New columns include event.counts which represents the sum total of values in the event column for each level of id or the sum total of levels of the state column if supplied as well as the count.weight column which sums the number of rows for each level of id.

Value

A data frame aggregated into count format.

Examples

# if the "event" column should be summed
long2count(long_data, id = "id", event = "event")
# if the "event" column contains levels that should be summed separately
long2count(long_data, id = "id", state = "event")

Long Format to Counting Process format

Description

Transforms data from long format to counting process format.

Usage

long2cp(data, id, time, status = NULL, drop = FALSE)

Arguments

data

A data frame with relevant columns.

id

A character string of the identification column name in data.

time

A character string of the time column name in data.

status

A character string of the status column in data either event or state.

drop

Logical indicator for whether any id groups with insufficient rows should be dropped from the output. Default is FALSE.

Details

The transition is primarily done by shifting the column supplied to the time argument into two new columns for a column-wise time definition and adjusting rows accordingly. Column names supplied to the status arguement are assumed to ocurr at the right endpoint so the first value for each id of the input is dropped. All other time-varying columns are assumed to ocurr at the left endpoint so the last value for each id of the input is dropped. The drop argument can be used for any id levels that may only have one row where a two column time data set might not suit them. Since there is not any useful gained from going from one time to the same time, it may be useful to just drop those id levels altogether.

Value

A data frame in counting process format.

Examples

long2cp(data = long_data, id = "id", time = "time", status = "event")

Long Format Data Example

Description

A toy data set in long format data.

Usage

long_data

Format

A data frame with 9 rows on the following 5 variables.

id

An identification variable

time

Time of observation

event

Status indicator variable

var1

First example explanatory variable

var2

Second example explanatory variable

Examples

long_data

Subset observations for grouped data based on first occurrence of a criteria value

Description

Takes all rows of a data frame up to and including the first occurrence of a supplied criteria for grouped data.

Usage

takefirst(data, id, criteria.column, criteria)

Arguments

data

A data frame with relevant columns.

id

A character string of the identification vector name defining groups in data.

criteria.column

The name as a character string of the column in data where the criteria is located.

criteria

The value of the cutoff for subsetting.

Details

Returns a data frame that takes all rows within the groups supplied by id up to and including the first occurrence of the value of criteria in criteria.column.

Value

A data frame subset up to and including the first row matching criteria in cirteria.column for each level of id.

Examples

takefirst(long_data, "id", criteria.column = "var1", criteria = 10.4)

Wide Format Data Example

Description

A toy data set in wide format.

Usage

wide_data

Format

A data frame with 3 rows on the following 14 variables.

id

An identification variable

time1

First time observation column

time2

Second time observation column

time3

Third time observation column

time4

Fourth observation column

event1

Status indicator at first time

event2

Status indicator at second time

event3

Status indicator at third time

event4

Status indicator at fourth time

var11

First explanatory variable at first time

var12

First explanatory variable at second time

var13

First explanatory variable at third time

var14

First explanatory variable at fourth time

var2

Second explanatory variable

Examples

wide_data