Type: Package
Title: Physical Activity with Accelerometers
Version: 1.2.3
Date: 2024-02-11
Author: Marco Geraci
Maintainer: Marco Geraci <marco.geraci@uniroma1.it>
Depends: R (≥ 3.0.0)
Imports: stats, utils, graphics, SparseM
Description: Functions to process, format and store ActiGraph GT1M and GT3X accelerometer data.
License: GPL-2 | GPL-3 [expanded from: GPL (≥ 2)]
LazyLoad: yes
NeedsCompilation: no
Packaged: 2024-02-11 15:27:20 UTC; marco
Repository: CRAN
Date/Publication: 2024-02-12 09:20:09 UTC

Physical Activity with Accelerometers

Description

This package provides processing and summary functions.

Details

Package: pawacc
Type: Package
Version: 1.2.3
Date: 2024-02-11
License: GPL (>= 2)
LazyLoad: yes

Author(s)

Marco Geraci

Maintainer: Marco Geraci <marco.geraci@uniroma1.it>

References

Geraci M, Rich C, Sera F, Cortina-Borja M, Griffiths LJ, and Dezateux C (2012). Technical report on accelerometry data processing in the Millennium Cohort Study. London, UK: University College London. Available at https://discovery.ucl.ac.uk/1361699

Aggregate Accelerometry Data

Description

This function aggregates count values

Usage

aggAccFile(object, by, which = "counts",
	x = NULL, keep.error = FALSE)

Arguments

object

an object of class accfile.

by

epoch by which count and steps are aggregated. Note: it cannot be less that the accelerometer epoch (object$info$epoch).

which

either 'counts' or 'steps' for gt1m files or one of c('x','y','z','steps') for gt3x files.

x

optional argument. If NULL, this is set to counts.

keep.error

logical flag. Should errors be omitted?

Value

outcome

aggregated values

ts_agg

time stamping

Author(s)

Marco Geraci

References

Geraci M, Rich C, Sera F, Cortina-Borja M, Griffiths LJ, and Dezateux C (2012). Technical report on accelerometry data processing in the Millennium Cohort Study. London, UK: University College London. Available at https://discovery.ucl.ac.uk/1361699

See Also

collapse.accfile

Examples


data(gt1m_sample)

# aggregate by 30-second epochs
aggAccFile(gt1m_sample, by = 30)

# aggregate by 5-minute epochs
aggAccFile(gt1m_sample, by = 300, keep.error = FALSE)

Collapse raw accelerometer files into a summary dataset

Description

This function collapses accelerometer data into a dataframe with summary statistics.

Usage

collapse(...)
## S3 method for class 'accfile'
collapse(object, which = "counts", palist = list(value = c(0, 100, 1000, 5000, 13000),
rescale.epoch = 60, labels = NULL, extreme = NULL), mwlist = list(value = 20,
nz = 0, rescale.epoch = 60), collapse.by = "%Y-%m-%d", collapse.epoch = 60, aggregate.by =
NULL, FUN.list = list(mean = function(x) mean(x, na.rm = TRUE)),
keep.extreme = FALSE, keep.error = FALSE, ...)

Arguments

object

an object of class gt1m.

which

either 'counts' or 'steps' for gt1m files or one of c('x','y','z','steps') for gt3x files.

palist

list of arguments for markpa.accfile.

mwlist

list of arguments for markwear.accfile.

collapse.by

dataset aggregation level. See argument format from strptime for options and details below.

collapse.epoch

epoch by which time spent in different physical activity modes is summarized. See details.

aggregate.by

pre-collapsing aggregation level for accelerometer values. See argument format from strptime for options.

FUN.list

a named list of functions. See fun.collapse.

keep.extreme

logical flag. If FALSE (default) extreme values will be replaced by NAs. See markpa.accfile.

keep.error

logical flag. If FALSE (default) data errors as identified by errorChk will be replaced by NAs.

...

arguments for dateSummary.

Value

A data frame containing the following variables

collapse.by

aggregation factor

fileid

file identifier

...

named columns according to arguments FUN and labels of palist

Author(s)

Marco Geraci

References

Geraci M, Rich C, Sera F, Cortina-Borja M, Griffiths LJ, and Dezateux C (2012). Technical report on accelerometry data processing in the Millennium Cohort Study. London, UK: University College London. Available at https://discovery.ucl.ac.uk/1361699

See Also

markpa.accfile, markwear.accfile, fun.collapse

Examples

## Not run: 

data(gt1m_sample)

collapse(gt1m_sample, palist = list(value = c(0, 100, 2000, 4000, 11000),
rescale.epoch = 60, labels = c("sedentary","light","moderate","vigorous","extreme_values"),
extreme = "last"), mwlist = list(value = 20,
nz = 0), collapse.by = "%Y-%m-%d", collapse.epoch = 60,
FUN.list = list(mean = function(x) round(mean(x, na.rm = TRUE),2)),
keep.extreme = FALSE, keep.error = FALSE)
$outcome
  collapse.by fileid outcome.mean sedentary  light moderate vigorous
1  2011-12-08   test       117.63    293.75 243.75    29.75     4.00
2  2011-12-09   test       157.83    349.75 143.25    33.50    24.50
3  2011-12-10   test        79.75    468.25 177.25    24.25     6.25
4  2011-12-11   test        57.96    355.50 126.00     8.50     3.25
5  2011-12-12   test        70.05    455.50 157.25    19.50     6.00
6  2011-12-13   test        72.99    475.50 181.50    15.25     8.00
7  2011-12-14   test        79.94    476.25 210.50    20.25     8.50
8  2011-12-15   test       232.50      0.00   0.50     0.00     0.00
9  2011-12-16   test          NaN      0.00   0.00     0.00     0.00
  extreme_values non-wear
1           0.00   388.75
2           6.00   883.00
3           0.50   763.50
4           0.25   946.50
5           0.00   801.75
6           0.25   759.50
7           1.50   723.00
8           0.00  1439.50
9           0.00   187.50

$call
collapse.accfile(object = gt1m_sample, palist = list(value = c(0, 
    100, 2000, 4000, 11000), rescale.epoch = 60, labels = c("sedentary", 
    "light", "moderate", "vigorous", "extreme_values"), extreme = "last"), 
    mwlist = list(value = 20, nz = 0), collapse.by = "
    collapse.epoch = 60, FUN.list = list(mean = function(x) round(mean(x, 
        na.rm = TRUE), 2)), keep.extreme = FALSE, keep.error = FALSE)

attr(,"class")
[1] "accfile.collapse"
Warning message:
In collapse.accfile(gt1m_sample, palist = list(value = c(0, 100,  :
  NAs imputed where extreme counts found

## End(Not run)

## Not run: 

collapse(gt1m_sample, palist = list(value = c(0, 100, 2000, 4000, 11000),
rescale.epoch = 60, labels = c("sedentary","light","moderate","vigorous","extreme_values"),
extreme = "last"), mwlist = list(value = 20,
nz = 0), collapse.by = "%Y-%m-%d", collapse.epoch = 60,
FUN.list = list(mean = function(x) round(mean(x, na.rm = TRUE),2),
sd = function(x) round(sd(x, na.rm = TRUE),2),
"95th" = function(x) round(quantile(x, probs = .95, na.rm = TRUE),2)),
keep.extreme = TRUE, keep.error = FALSE)
$outcome
  collapse.by fileid outcome.mean outcome.sd outcome.95th sedentary  light
1  2011-12-08   test       117.63     216.12       529.40    293.75 243.75
2  2011-12-09   test       201.10     567.65      1085.60    349.75 143.25
3  2011-12-10   test        81.97     221.33       465.50    468.25 177.25
4  2011-12-11   test        59.80     172.08       320.35    355.50 126.00
5  2011-12-12   test        70.05     188.49       401.00    455.50 157.25
6  2011-12-13   test        74.08     207.55       386.85    475.50 181.50
7  2011-12-14   test        87.42     275.27       415.30    476.25 210.50
8  2011-12-15   test       232.50     222.74       374.25      0.00   0.50
9  2011-12-16   test          NaN         NA           NA      0.00   0.00
  moderate vigorous extreme_values non-wear
1    29.75     4.00           0.00   388.75
2    33.50    24.50           6.00   883.00
3    24.25     6.25           0.50   763.50
4     8.50     3.25           0.25   946.50
5    19.50     6.00           0.00   801.75
6    15.25     8.00           0.25   759.50
7    20.25     8.50           1.50   723.00
8     0.00     0.00           0.00  1439.50
9     0.00     0.00           0.00   187.50

$call
collapse.accfile(object = gt1m_sample, palist = list(value = c(0, 
    100, 2000, 4000, 11000), rescale.epoch = 60, labels = c("sedentary", 
    "light", "moderate", "vigorous", "extreme_values"), extreme = "last"), 
    mwlist = list(value = 20, nz = 0), collapse.by = "
    collapse.epoch = 60, FUN.list = list(mean = function(x) round(mean(x, 
        na.rm = TRUE), 2), sd = function(x) round(sd(x, na.rm = TRUE), 
        2), `95th` = function(x) round(quantile(x, probs = 0.95, 
        na.rm = TRUE), 2)), keep.extreme = TRUE, keep.error = FALSE)

attr(,"class")
[1] "accfile.collapse"

## End(Not run)

Date summary for accelerometer files

Description

This function provides a date summary for Actigraph GT1M accelerometer files.

Usage

dateSummary(object, wear, timestamp, minval = 0,
rescale.epoch = 60, keep.error = FALSE)

Arguments

object

an object of class gt1m.

wear

a vector that classifies wear and non-wear time. See markwear.accfile.

timestamp

a timestamp vector for accelerometer values that can be provided by tsFormat.

minval

threshold defining the minimum number of minutes to identify first and last days. See details.

rescale.epoch

epoch expressed in the same unit as acceleromenter's epoch to determine minutes of wear time (default is 60 and assumed to be in seconds).

keep.error

logical flag. If FALSE (default) data errors as identified by errorChk will be replaced by NAs.

Details

Based on total wear time (in minutes) for each day, the theshold minval is applied to identify the first and last days. For example, if accelerometers are sent by post to collect survey data, the first and last days in which the accelerometer was worn might not be known. Days before the first and those after the last day (truncated days) are discarded by collapse.accfile.

Value

A data frame containing the following variables

fileid

file identifier

days

dates by calendar day

freq

frequency of accelerometer observations in each day

hour_day

total hours of accelerometer observations in each day

start_day

starting time of accelerometer observations

end

end time of accelerometer observations

valid_mins

wear time (minutes)

IsStartDate

dummy variable to define starting date (1 = yes)

IsEndDate

dummy variable to define end date (1 = yes)

IsTruncated

dummy variable to define truncated date (1 = yes)

Author(s)

Marco Geraci

References

Geraci M, Rich C, Sera F, Cortina-Borja M, Griffiths LJ, and Dezateux C (2012). Technical report on accelerometry data processing in the Millennium Cohort Study. London, UK: University College London. Available at https://discovery.ucl.ac.uk/1361699

See Also

collapse.accfile

Examples

data(gt1m_sample)

mw <- markwear.accfile(gt1m_sample, 20)
tsf <- tsFormat(gt1m_sample)
dateSummary(gt1m_sample, mw, tsf)
## Not run: 
  fileid       days freq hour_day start_day end_day valid_mins IsStartDate
1   test 2011-12-08 3840   16.000        08      23     571.25           1
2   test 2011-12-09 5760   24.000        00      23     557.00           0
3   test 2011-12-10 5760   24.000        00      23     676.50           0
4   test 2011-12-11 5760   24.000        00      23     493.50           0
5   test 2011-12-12 5760   24.000        00      23     638.25           0
6   test 2011-12-13 5760   24.000        00      23     680.50           0
7   test 2011-12-14 5760   24.000        00      23     717.00           0
8   test 2011-12-15 5760   24.000        00      23       0.50           0
9   test 2011-12-16  750    3.125        00      03       0.00           0
  IsEndDate IsTruncated
1         0           0
2         0           0
3         0           0
4         0           0
5         0           0
6         0           0
7         0           0
8         0           0
9         1           0

## End(Not run)

# at least 600 minutes per day to determine first and last day
dateSummary(gt1m_sample, mw, tsf, minval = 600)
## Not run: 
  fileid       days freq hour_day start_day end_day valid_mins IsStartDate
1   test 2011-12-08 3840   16.000        08      23     571.25           0
2   test 2011-12-09 5760   24.000        00      23     557.00           0
3   test 2011-12-10 5760   24.000        00      23     676.50           1
4   test 2011-12-11 5760   24.000        00      23     493.50           0
5   test 2011-12-12 5760   24.000        00      23     638.25           0
6   test 2011-12-13 5760   24.000        00      23     680.50           0
7   test 2011-12-14 5760   24.000        00      23     717.00           0
8   test 2011-12-15 5760   24.000        00      23       0.50           0
9   test 2011-12-16  750    3.125        00      03       0.00           0
  IsEndDate IsTruncated
1         0           1
2         0           1
3         0           0
4         0           0
5         0           0
6         0           0
7         1           0
8         0           1
9         0           1

## End(Not run)

Perform error checking

Description

These functions look for errors in the data. A code is returned.

Usage

errorChk(x, fault = 32767)

Arguments

x

vector of accelerometer data.

fault

numerical value that indicates voltage signal saturation (temporarily used for both accelerometer counts and steps).

Details

Error coded are as follow: 0, no error; 1, all values are 5-digit values or all one value; 2, negative values; 3, NAs.

Value

a vector of the same length as x.

Author(s)

Marco Geraci

References

Geraci M, Rich C, Sera F, Cortina-Borja M, Griffiths LJ, and Dezateux C (2012). Technical report on accelerometry data processing in the Millennium Cohort Study. London, UK: University College London. Available at https://discovery.ucl.ac.uk/1361699

See Also

gt1mAccFile, gt3xAccFile

Summary statistics

Description

Accessory function for collapsing accelerometer files.

Usage

fun.collapse(x, fun = list(mean = function(x) mean(x, na.rm = TRUE),
median = function(x) median(x, na.rm = TRUE),
sd = function(x) sd(x, na.rm = TRUE)))

Arguments

x

numeric vector.

fun

named list of functions to be applied to x.

Value

a list of named values of the same length as fun.

Author(s)

Marco Geraci

References

Geraci M, Rich C, Sera F, Cortina-Borja M, Griffiths LJ, and Dezateux C (2012). Technical report on accelerometry data processing in the Millennium Cohort Study. London, UK: University College London. Available at https://discovery.ucl.ac.uk/1361699

See Also

collapse.accfile

Read list of accelerometer files

Description

This function reads a list of Actigraph GT1M accelerometer files.

Usage

gt1mAccDir(accFileList, save, compress = "gzip",
	compression_level = 6, progbar = TRUE)

Arguments

accFileList

an object of type list.

save

either logical or character. See readAccDir for details.

compress

logical or character string specifying whether saving to a named file is to use compression if save = TRUE. See argument compress in save.

compression_level

integer: the level of compression to be used. See argument compression_level in save.

progbar

logical flag. Should a progress bar be used? Available for Windows only.

Value

An object of class acclist.

Author(s)

Marco Geraci

References

Actigraph (Pensacola, Florida).

Geraci M, Rich C, Sera F, Cortina-Borja M, Griffiths LJ, and Dezateux C (2012). Technical report on accelerometry data processing in the Millennium Cohort Study. London, UK: University College London. Available at https://discovery.ucl.ac.uk/1361699

See Also

readAccDir

Read a single accelerometer file

Description

This function reads an Actigraph GT1M accelerometer file.

Usage

gt1mAccFile(file, path, fileid, counts.pos = 1,
	tz = "Europe/London", sparse = FALSE,
	fault = 32767)

Arguments

file

file name including file extension.

path

path to file.

fileid

label for file identifier.

counts.pos

when storage mode allows for accelerometer counts and steps to be recorded at the same time, this argument specifies the position of first measurement of accelerometer counts (default is counts.pos = 1).

tz

a character string specifying the timezone to be used for the conversion (see strptime).

sparse

logical flag: should data be stored in sparse format?

fault

numerical value that indicates voltage signal saturation.

Details

Raw accelerometer data are processed according to the device data format. Several data checks are performed by errorChk and infoDate. An additional check is performed on the lenght of the sequence of measurements when both accelerometer counts and steps are recorded. If the length is odd, a warning message is produced. See file 'gt1m_sample.dat' in directory '\inst\extdata' of this package.

Value

These functions return an object of two classes: accfile and additional device-specific class (i.e., gt1m).

An object of class accfile is a list containing the following components:

df

A data.frame object with accelerometer values in columns counts and steps (if present), and coded error for each accelerometer data column. See errorChk for error codes. If sparse = TRUE, all variables of the data frame df are returned as vectors of a matrix in sparse format (see as.matrix.csr for details).

info

A data.frame object with file identifier (fileid), device serial number (serial), number of recorded measurements (nobs), epoch (epoch), accelerometer mode (mode), start date and time (ts_start), time zone (tz), battery voltage (voltage), download date and time (ts_dl).

error_summary

A list object with file identifier (fileid), summary tables of error codes for each accelerometer data column, error code for date (date), and logical flag for odd number of measurements (odd_number) (see details).

Author(s)

Marco Geraci

References

Actigraph (Pensacola, Florida).

Geraci M, Rich C, Sera F, Cortina-Borja M, Griffiths LJ, and Dezateux C (2012). Technical report on accelerometry data processing in the Millennium Cohort Study. London, UK: University College London. Available at https://discovery.ucl.ac.uk/1361699

See Also

readAccDir, gt1mAccDir

Examples


data(gt1m_sample)
class(gt1m_sample)

GT1M accelerometer file

Description

This is the output of gt1mAccFile.

Format

See value in gt1mAccFile. The 'raw' file 'gt1m_sample.dat' can be found in the directory '\inst\extdata' of this package.

Source

Actigraph (Pensacola, Florida).

Examples

## Not run: 
gt1m_sample <- gt1mAccFile("gt1m_sample.dat", path = "", fileid = "test")
gt1m_sample
  fileid serial  nobs epoch mode            ts_start  tz voltage               ts_dl
1   test  xxxxx 44910    15    1 2011-12-08 08:00:00 GMT    3.76 2011-12-24 11:20:36

## End(Not run)

Read list of accelerometer files

Description

This function reads a list of Actigraph GT3X accelerometer files.

Usage

gt3xAccDir(accFileList, save, compress = "gzip",
	compression_level = 6, progbar = TRUE)

Arguments

accFileList

an object of type list.

save

either logical or character. See readAccDir for details.

compress

logical or character string specifying whether saving to a named file is to use compression if save = TRUE. See argument compress in save.

compression_level

integer: the level of compression to be used. See argument compression_level in save.

progbar

logical flag. Should a progress bar be used? Available for Windows only.

Value

An object of class acclist.

Author(s)

Marco Geraci

References

Actigraph (Pensacola, Florida).

See Also

readAccDir

Read a single accelerometer file

Description

This function reads Actigraph GT3X and ActiSleep accelerometer files.

Usage

gt3xAccFile(file, path, fileid, tz = "Europe/London",
	sparse = FALSE, fault = 32767)

Arguments

file

file name including file extension.

path

path to file.

fileid

label for file identifier.

tz

a character string specifying the timezone to be used for the conversion (see strptime).

sparse

logical flag: should data be stored in sparse format?

fault

numerical value that indicates voltage signal saturation.

Details

Raw accelerometer data are processed according to the device data format. See file 'gt3x_sample.dat' in directory '\inst\extdata' of this package.

Value

These functions return an object of two classes: accfile and additional device-specific class (i.e., gt3x).

An object of class accfile is a list containing the following components:

df

A data.frame object with accelerometer values in columns y, x, z, and steps (if present), and coded error for each accelerometer data column. See errorChk for error codes. If sparse = TRUE, all variables of the data frame df are returned as vectors of a matrix in sparse format (see as.matrix.csr for details).

info

A data.frame object with file identifier (fileid), device serial number (serial), number of recorded measurements (nobs), epoch (epoch), accelerometer mode (mode), start date and time (ts_start), time zone (tz), battery voltage (voltage), download date and time (ts_dl).

error_summary

A list object with file identifier (fileid), summary tables of error codes for each accelerometer data column, error code for date (date), and logical flag for odd number of measurements (odd_number) (see details).

Author(s)

Marco Geraci

References

Actigraph (Pensacola, Florida).

Geraci M, Rich C, Sera F, Cortina-Borja M, Griffiths LJ, and Dezateux C (2012). Technical report on accelerometry data processing in the Millennium Cohort Study. London, UK: University College London. Available at https://discovery.ucl.ac.uk/1361699

See Also

readAccDir

Examples


data(gt3x_sample)
class(gt3x_sample)

GT3X accelerometer file

Description

This is the output of gt3xAccFile.

Format

See value in gt3xAccFile. The 'raw' file 'gt3x_sample.dat' can be found in the directory '\inst\extdata' of this package.

Source

Actigraph (Pensacola, Florida).

Examples

## Not run: 
gt3x_sample <- gt3xAccFile("gt3x_sample.dat", path = "", fileid = "test")
gt3x_sample
  fileid serial nobs epoch mode            ts_start  tz voltage               ts_dl
1   test  xxxxx 2676     1   13 2009-03-03 10:40:00 GMT    4.09 2009-03-03 11:24:49

## End(Not run)

Classify accumulation of physical activity in bouts

Description

This function identifies bouts of physical activity using user-defined breakpoints for accelerometer counts.

Usage

markbouts(object, value, which = "counts", bts = c(0, 10, 20, Inf), rescale.epoch = 60,
collapse.by = "%Y-%m-%d", value.labels = NULL, bouts.labels = NULL,
extreme = NULL, keep.error = FALSE, progbar = TRUE)
## S3 method for class 'acclist'
markbouts(object, value, which = "counts", bts = c(0, 10, 20, Inf), rescale.epoch = 60,
collapse.by = "%Y-%m-%d", value.labels = NULL, bouts.labels = NULL,
extreme = NULL, keep.error = FALSE, progbar = TRUE)
## S3 method for class 'accfile'
markbouts(object, value, which = "counts", bts = c(0, 10, 20, Inf), rescale.epoch = 60,
collapse.by = "%Y-%m-%d", value.labels = NULL, bouts.labels = NULL,
extreme = NULL, keep.error = FALSE, progbar = NULL)

Arguments

object

an object of class acclist or accfile.

value

vector of breaks to define physical activity modes. The lowest breakpoint must be 0.

which

either 'counts' or 'steps' for gt1m files or one of c('x','y','z','steps') for gt3x files.

bts

vector of breaks to define bouts duration (in minutes).

rescale.epoch

epoch expressed in the same unit as acceleromenter's epoch to determine multiplier to rescale value (default is 60). See details.

collapse.by

dataset aggregation level. See argument format from strptime for options and details below.

value.labels

labels for physical activity modes.

bouts.labels

labels for bouts duration catergories (NULL is recommended).

extreme

if value includes a threshold for defining the category of extreme values, this argument identifies such category among the physical activity modes defined by value. See details.

keep.error

logical flag. If FALSE (default) data errors as identified by errorChk will be replaced by NAs.

progbar

logical flag. Should a progress bar be used? Available for Windows only. Argument not used for class accfile.

Details

Breakpoints are specified as counts per seconds using value. If the epoch used for value is different from the accelerometer's epoch, a rescaling is applied. E.g., if epoch is 15 seconds and breakpoints are expressed as counts per 60 seconds, value is divided by 60/15 = 4. There can be n physical activity modes at maximum, where n is the length of value.

The argument extreme is NULL by default. Use either ‘last’ to select the last category or the category number 1 to n.

Value

The function markbouts.accfile returns duration, frequency and mean duration of bouts by bout category, physical activity mode and T levels of collapse.by (e.g., day) in array of dimension c(length(bts) - 1,length(value) + 1,3,T). markbouts.acclist is applied to objects of class acclist, in which case a list of arrays of the same length as the number of accelerometer files in object is returned.

Author(s)

Marco Geraci

See Also

readAccDir

Classify mode of physical activity

Description

This function identifies modes of physical activity using user-defined breakpoints for accelerometer counts.

Usage

markpa(object, value, which = "counts", rescale.epoch = 60, labels = NULL,
extreme = NULL, keep.error = FALSE, progbar = TRUE)
## S3 method for class 'acclist'
markpa(object, value, which = "counts", rescale.epoch = 60, labels = NULL,
extreme = NULL, keep.error = FALSE, progbar = TRUE)
## S3 method for class 'accfile'
markpa(object, value, which = "counts", rescale.epoch = 60, labels = NULL,
extreme = NULL, keep.error = FALSE, progbar = NULL)

Arguments

object

an object of class acclist or accfile.

value

vector of breaks to define physical activity modes. The lowest breakpoint must be 0.

which

either 'counts' or 'steps' for gt1m files or one of c('x','y','z','steps') for gt3x files.

rescale.epoch

epoch expressed in the same unit as acceleromenter epoch to determine multiplier to rescale value (default is 60). See details.

labels

labels for physical activity modes.

extreme

if value includes a threshold for defining the category of extreme values, this argument identifies such category among the physical activity modes defined by value. See details.

keep.error

logical flag. If FALSE (default) data errors as identified by errorChk will be replaced by NAs.

progbar

logical flag. Should a progress bar be used? Available for Windows only. Argument not used for class accfile.

Details

Breakpoints are specified as counts per seconds using value. If the epoch used for value is different from the accelerometer epoch, a rescaling is applied. E.g., if epoch is 15 seconds and breakpoints are expressed as counts per 60 seconds, value is divided by 60/15 = 4. There can be n physical activity modes at maximum, where n is the length of value.

The argument extreme is NULL by default. Use either ‘last’ to select the last category or the category number 1 to n.

Value

If object is of class acclist, a list of factors of the same length as the number of accelerometer files in object. If object is of class accfile, a single factor will be given. The number of levels is equal to length(value) + 1.

Author(s)

Marco Geraci

See Also

readAccDir

Classify wear and non-wear time

Description

This functions identifies sequences of zeroes of a given length to classify wear and non-wear time in accelerometer data files.

Usage

markwear(object, value, which = "counts", rescale.epoch = 60,
	nz = 0, keep.error = FALSE, progbar = TRUE)
## S3 method for class 'acclist'
markwear(object, value, which = "counts", rescale.epoch = 60,
	nz = 0, keep.error = FALSE, progbar = TRUE)
## S3 method for class 'accfile'
markwear(object, value, which = "counts", rescale.epoch = 60,
	nz = 0, keep.error = FALSE, progbar = NULL)

Arguments

object

an object of class acclist or accfile.

value

the length of the time window (in minutes) which contains zero values.

which

either 'counts' or 'steps' for gt1m files or one of c('x','y','z','steps') for gt3x files.

rescale.epoch

epoch expressed in the same unit as acceleromenter epoch to determine multiplier to rescale value (default is 60). See details.

nz

the length of the time window (in minutes) of non-zero value sequences allowed between every two sequences of zero values.

keep.error

logical flag. If FALSE (default) data errors as identified by errorChk will be replaced by NAs.

progbar

logical flag. Should a progress bar be used? Available for Windows only. Argument not used for class accfile.

Details

The accelerometer epoch is assumed to be expressed in seconds. Therefore value is automatically rescaled to value * 60/object$info$epoch.

Value

If object is of class acclist, a list of factors with two levels of the same length as the number of accelerometers files in object. If object is of class accfile, a single factor will be given. Levels are 'non-wear' and 'wear'.

Author(s)

Marco Geraci

References

Geraci M, Rich C, Sera F, Cortina-Borja M, Griffiths LJ, and Dezateux C (2012). Technical report on accelerometry data processing in the Millennium Cohort Study. London, UK: University College London. Available at https://discovery.ucl.ac.uk/1361699

See Also

readAccDir

Examples


## Not run: 

data(gt1m_sample)

# 20-minute threshold
wear <- markwear(gt1m_sample, value = 20)
> table(wear)
 wear
 Non-wear     Wear 
    27572    17338 

## End(Not run)

## Not run: 

# lower threshold
wear <- markwear(gt1m_sample, value = 5)
> table(wear)
 wear
 Non-wear     Wear 
    30188    14722 

## End(Not run)

## Not run: 

# allow for some non-zero values within a 20-minute window
wear <- markwear.accfile(gt1m_sample, value = 20, nz = 2)
> table(wear)
 wear
 Non-wear     Wear 
   28198    16712 

## End(Not run)

Internal pawacc objects

Description

Internal pawacc objects.

Details

These are not to be called by the user.

Plot accelerometer file

Description

Thes functions plot data from Actigraph GT1M and GT3X accelerometer files.

Usage

## S3 method for class 'gt1m'
plot(x, y = NULL, xlab, ylab, main,
keep.error = TRUE, which = "counts", select = 1,...)
## S3 method for class 'gt3x'
plot(x, y = NULL, xlab, ylab, main,
keep.error = TRUE, which = "x", select = 1,...)

Arguments

x

an object of class gt1m. It can be either accfile or acclist.

y

ignored.

xlab

x-axis label (optional).

ylab

y-axis label (optional).

main

main title (optional).

keep.error

logical flag. If FALSE (default) data errors as identified by errorChk will be replaced by NAs.

which

either 'counts' or 'steps' for gt1m files or one of c('x','y','z','steps') for gt3x files.

select

numeric. If class(x) is acclist, this argument specifies the corresponding position of the accelerometer file in the list (first file by default).

...

Arguments to be passed to methods, such as graphical parameters (see par).

Author(s)

Marco Geraci

References

Actigraph (Pensacola, Florida).

Geraci M, Rich C, Sera F, Cortina-Borja M, Griffiths LJ, and Dezateux C (2012). Technical report on accelerometry data processing in the Millennium Cohort Study. London, UK: University College London. Available at https://discovery.ucl.ac.uk/1361699

See Also

plot, par, gt1mAccFile, gt1mAccDir

Examples

data(gt1m_sample)
data(gt3x_sample)


plot(gt1m_sample, which = "counts")
plot(gt1m_sample, which = "steps")
plot(gt3x_sample, which = "x")
plot(gt3x_sample, which = "steps")

Print an accfile or acclist Object

Description

Print an object generated by gt1mAccFile, gt3xAccFile or readAccDir.

Usage

## S3 method for class 'acclist'
print(x, ...)
## S3 method for class 'accfile'
print(x, ...)

Arguments

x

an accfile or an acclist object.

...

not used.

Author(s)

Marco Geraci

Read content of a directory with accelerometer files

Description

This function reads raw files downloaded from accelerometers and stores them in one directory.

Usage

readAccDir(path, model, ext = "dat", counts.pos = 1,
	tz = "Europe/London", sparse = FALSE,
	fault = 32767, save = TRUE, compress = "gzip",
	compression_level = 6,...)

Arguments

path

path to accelerometer files.

model

accelerometer model, either 'gt1m' or 'gt3x'.

ext

file extension without '.' (default is 'dat').

counts.pos

position of first measurement of accelerometer counts. See gt1mAccFile for details.

tz

a character string specifying the timezone to be used for the conversion (see strptime).

sparse

logical flag: should data be stored in sparse format?

fault

numerical value that indicates voltage signal saturation (temporarily used for both accelerometer counts and steps).

save

either logical or character. If logical, save = TRUE stores accelerometer files as .Rdata objects in a new directory ('accRfiles') in path. If the directory already exists, confirmation for overwriting will be prompted. save = FALSE stores accelerometer files in current R session. Alternatively, an existing folder can be specified.

compress

logical or character string specifying whether saving to a named file is to use compression if save = TRUE. See argument compress in save.

compression_level

integer: the level of compression to be used. See argument compression_level in save.

...

not used.

Details

This is a wrapper function. It reads raw accelerometer files stored in one directory. The argument ext specifies the extension of the files to look for (e.g., 'dat' or 'txt'). Files with different extension and/or sub-directories are ignored. The data format must be consistent with the specification of model.

Value

Either a set of .Rdata files or a list of objects of class accfile. In both cases, each accelerometer data file is stored as an object of type list and labelled using the data file name. See gt1mAccFile or gt3xAccFile for details.

Author(s)

Marco Geraci

References

Geraci M, Rich C, Sera F, Cortina-Borja M, Griffiths LJ, and Dezateux C (2012). Technical report on accelerometry data processing in the Millennium Cohort Study. London, UK: University College London. Available at https://discovery.ucl.ac.uk/1361699

See Also

gt1mAccFile, gt3xAccFile, gt1mAccDir, gt3xAccDir

Time Stamping

Description

Time stamping.

Usage

tsFormat(object)

Arguments

object

an object of class accfile.

Value

a vector of timestamps.

Author(s)

Marco Geraci

See Also

tsFromEpoch

Examples


data(gt1m_sample)
data(gt3x_sample)

tsFormat(gt1m_sample)
tsFormat(gt3x_sample)

Calculate timestamp from epoch number or epoch number from timestamp

Description

Utility functions.

Usage

tsFromEpoch(object, x)
epochFromTS(object, x)

Arguments

object

an object of class accfile.

x

either an integer giving the epoch number or the timestamp in a POSIX format (e.g., '%Y-%m-%d %H:%M:%S').

Value

either a timestamp corresponding to an epoch number or the epoch number corresponding to a timestamp.

Author(s)

Marco Geraci

See Also

gt1mAccFile

Examples

data(gt1m_sample)

tsFromEpoch(gt1m_sample, 10000)
# [1] "2011-12-10 01:39:45 GMT"

epochFromTS(gt1m_sample, as.POSIXlt(strptime('2011-12-10 01:39:45', '%Y-%m-%d %H:%M:%S')))
# [1] 10000