`walkboutr`

The goal of walkboutr is to process GPS and accelerometry data into walking bouts. walkboutr will either return the original dataset along with bout labels and categories, or a summarized and de-identified dataset that can be shared for collaboration.

Installation

You can install the development version of walkboutr from GitHub with:

# install.packages("devtools")
devtools::install_github("rwalkbout/walkboutr")

Basic Usage

Simulated sample data

This is an example of simulated data that could be processed by walkboutr. The GPS data contain the required columns: time, latitude, longitude, speed. The accelerometry data contain the required columns: time, accerometry counts. These data have no extra columns, do not contain NAs, and don’t contain negative speeds or accelerometry counts. All times are also in date-time format.

library(walkboutr)
# generate sample gps data:
gps_data <- generate_walking_in_seattle_gps_data() 
# generate sample accelerometry data:
accelerometry_counts <- make_full_day_bout_without_metadata()

GPS data:

time	latitude	longitude	speed
2012-04-07 00:00:30	47.60620	122.3321	1.5510573
2012-04-07 00:01:00	47.61333	122.3392	2.4004880
2012-04-07 00:01:30	47.62225	122.3482	0.6412646
2012-04-07 00:02:00	47.62541	122.3513	1.6616599
2012-04-07 00:02:30	47.63164	122.3575	2.0068013
2012-04-07 00:03:00	47.63945	122.3654	1.1009735

Accelerometry data:

activity_counts	time
0	2012-04-07 00:00:30
0	2012-04-07 00:01:00
0	2012-04-07 00:01:30
0	2012-04-07 00:02:00
500	2012-04-07 00:02:30
500	2012-04-07 00:03:00

Now that we have sample data, we can look at how the walkboutr package works. There are two top level functions that will allow us to generate either (1) a dataset with bouts and bout categories with all of our original data included, or (2) a summary dataset that is completely de-identified and shareable for research purposes.

Walk bout dataset including original data

walk_bouts <- identify_walk_bouts_in_gps_and_accelerometry_data(gps_data,accelerometry_counts)

bout	bout_category	activity_counts	time	non_wearing	complete_day	latitude	longitude	speed
1	walk_bout	500	2012-04-07 00:03:00	FALSE	TRUE	47.63945	122.3654	1.1009735
1	walk_bout	500	2012-04-07 00:05:00	FALSE	TRUE	47.68225	122.4082	2.7901428
1	walk_bout	500	2012-04-07 00:07:00	FALSE	TRUE	47.74325	122.4692	0.9801357
1	walk_bout	500	2012-04-07 00:05:30	FALSE	TRUE	47.69541	122.4213	2.7249735
1	walk_bout	500	2012-04-07 00:06:00	FALSE	TRUE	47.70804	122.4340	4.0867381
1	walk_bout	500	2012-04-07 00:06:30	FALSE	TRUE	47.72800	122.4539	3.0513150

Summarized walk bout dataset

This dataset is a set of labelled bouts that are categorized (bout_category) and contains information on bout specific median speed (median_speed), the start time of the bout (bout_start), the duration of the bout (in minutes for computational ease, duration), and a flag for whether the bout came from a dataset with a complete day worth of data (complete_day).

summary <- summarize_walk_bouts(walk_bouts)

bout	median_speed	complete_day	bout_start	duration	bout_category
1	2.736466	TRUE	2012-04-07 00:02:30	5.0	walk_bout
2	2.555720	TRUE	2012-04-07 00:09:30	4304.5	walk_bout

In this example, we have 2 bout(s), and each bout has a label. Bout 1 occurred on 2012.04.07 and has a complete day worth of data (complete_day = TRUE) and a start time of 00:02:30. This bout lasted 5 minutes, or 0.0833333 hours. The bout is a non-walk bout because the participant was moving too slowly for this walk to be considered walking.

For more information on bout categories and how these are assigned, please see the vignette titled Generate Walk Bouts.