README.md

rwalkabout

rwalkbout aims to streamline the identification of walk bouts from accelerometry and gps data.

Set up

Set working directory to be rwalkabout/R.

Edit and run rwalkabout/analysis/analyze_all_subject_seattle.R.

Customize rwalkbout to your dataset

rwalkbout can be configured to any gps and accelerometry (acc) data, as long as you write the respective acc/gps file reader.

The reader should be a function that reads the data and return to rwalkbout with a predefined format.

Below outlines what inputs rwalkbout will give to the function and what returned format rwalkbout expects.

Note: ensure that date_time returned by both readers are in the same time zone

File reader for acc data

Inputs:

accepts a parameter acc_file_path

Outputs:

returns a dataframe with columns date_time,count

cast to tibble with as.tibble()

collapse/interpolate to make epoch period=15s

Example outputs and data types:

> acc_data$date_time[1]
[1] "2016-12-10 00:00:15 PST"
> class(acc_data$date_time[1])
[1] "POSIXct" "POSIXt" 

> acc_data$count[1]
[1] 251
> class(acc_data$count)
[1] "numeric"

File reader for gps data

Inputs:

accepts a parameter gps_file_path

Outputs:

returns a dataframe with columns date_time,latitude,longitude,speed, latitude and longitude should be positive/negative based its N/S, W/E

cast to tibble with as.tibble()

Example outputs and data types:

> class(gps_data$date_time[1])
[1] "POSIXct" "POSIXt" 
> gps_data$date_time[1]
[1] "2016-12-10 03:37:05 PST"

> class(gps_data$latitude[1])
[1] "numeric"
> gps_data$latitude[1]
[1] 45.437084

> class(gps_data$longitude[1])
[1] "numeric"
> gps_data$longitude[1]
[1] -122.62609

> class(gps_data$speed[1])
[1] "numeric"
> gps_data$speed[1]
[1] 0.590584

acc and gps file mapper

Optionally users can create a function that returns the mapping between acc and gps files. Examples are shown in rwalkbout/R/subject_file_mapper.R.

rwalkbout does not require it for running. It is just a nice thing to do when dealing with folders of files and subjects.