Visiting Shrewsbury
How many times did I visit Shrewsbury? Or at least how many times did my telephone visit Shrewsbury?
This is based on this strategy https://shiring.github.io/maps/2016/12/30/Standortverlauf_post for querying the location data.
Google Take Out https://takeout.google.com/settings/takeout provides my location data in Json format, with jsonlite will load. The Location History.json is already stripped down to 7 days, so there is less of it to handle. The whole file is pretty big so is difficult to work with on Azure Notebooks.
# Load the package
library(jsonlite)
system.time(import <- fromJSON("Location History.json"))
user system elapsed
0.882 0.011 1.286
# extracting the locations dataframe
locations = import$locations
head(locations[,c("latitudeE7", "longitudeE7", "timestampMs")], n=3)
| latitudeE7 | longitudeE7 | timestampMs |
|---|---|---|
| 525622196 | -27951121 | 1541616455811 |
| 525622196 | -27951121 | 1541616049051 |
| 525622196 | -27951121 | 1541615881157 |
Converting the Data
The dates are held as POSIX dates (milliseconds since the begining of time) and need to be converted to to queried in a readable manner. Also the locations are held as integers (for convenience) so need to be divided by 10^7 to get the latitude and longitudes. The ‘E7’ on the field implies this.
# converting time column from posix milliseconds into a readable time scale
locations$date_time = as.POSIXct(as.numeric(locations$timestampMs)/1000, origin = "1970-01-01")
locations$date <- as.Date(locations$date_time)
# converting longitude and latitude from E7 to GPS coordinates
locations$lat = locations$latitudeE7 / 1e7
locations$lon = locations$longitudeE7 / 1e7
# To get the days of the week so I can query for thing 'weekend' install and load the package `lubridate`
install.packages("lubridate")
library(lubridate)
locations$weekday <- wday(as.Date(locations$date_time))
head(locations[,c("lat", "lon", "date", "weekday")], n=3)
Installing package into ‘/home/nbuser/R’
(as ‘lib’ is unspecified)
Warning message in install.packages("lubridate"):
“installation of package ‘lubridate’ had non-zero exit status”
Attaching package: ‘lubridate’
The following object is masked from ‘package:base’:
date
| lat | lon | date | weekday |
|---|---|---|---|
| 52.56222 | -2.795112 | 2018-11-07 | 4 |
| 52.56222 | -2.795112 | 2018-11-07 | 4 |
| 52.56222 | -2.795112 | 2018-11-07 | 4 |
Select the Visits
Using a box around Shrewsbury where:
- bottom left = 52.704117, -2.762084
- top right = 52.709638, -2.748738
And selecting the where the weekday is between 2 and 6 (Sunday = 1 and Saturday = 7, don’t ask me why).
locations_in_shrewsbury <- subset(locations, lat > 52.704117 & lat < 52.709638 & lon< -2.748738 & lon> -2.762084 & weekday > 1 & weekday < 7)
visits_to_shrewsbury_in_working_week <- unique(locations_in_shrewsbury[c("date","weekday")])
head(visits_to_shrewsbury_in_working_week, n=10)
| date | weekday | |
|---|---|---|
| 66 | 2018-11-07 | 4 |
| 451 | 2018-11-06 | 3 |
| 983 | 2018-11-02 | 6 |
Giving 3 visits to Shrewsbury in those 7 days.