Any reason for not using `sf::read_sf()`

[ColinFay]

Hey,

I was reading the code from process_data() and was wondering why this function was not relying on sf::read_sf() ?

As far as I can tell it gives a more straightforward interface and should be more reliable:

> sf::read_sf("inst/gpx/running/activity_7371775190.gpx", layer = "track_points")
Simple feature collection with 527 features and 27 fields
Geometry type: POINT
Dimension:     XY
Bounding box:  xmin: -1.275479 ymin: 51.38637 xmax: -1.252118 ymax: 51.40214
Geodetic CRS:  WGS 84
# A tibble: 527 × 28
   track_fid track_seg_id track_seg_point_id   ele time                magvar
       <int>        <int>              <int> <dbl> <dttm>               <dbl>
 1         0            0                  0 -46.2 2021-08-26 08:22:57     NA
 2         0            0                  1 -46   2021-08-26 08:22:58     NA
 3         0            0                  2 -45.8 2021-08-26 08:23:01     NA
 4         0            0                  3 -46.2 2021-08-26 08:23:07     NA
 5         0            0                  4 -46.8 2021-08-26 08:23:14     NA
 6         0            0                  5 -46.6 2021-08-26 08:23:20     NA
 7         0            0                  6 -46.6 2021-08-26 08:23:26     NA
 8         0            0                  7 -46.6 2021-08-26 08:23:27     NA
 9         0            0                  8 -46.8 2021-08-26 08:23:34     NA
10         0            0                  9 -46.6 2021-08-26 08:23:36     NA
# … with 517 more rows, and 22 more variables: geoidheight <dbl>, name <chr>,
#   cmt <chr>, desc <chr>, src <chr>, link1_href <chr>, link1_text <chr>,
#   link1_type <chr>, link2_href <chr>, link2_text <chr>, link2_type <chr>,
#   sym <chr>, type <chr>, fix <chr>, sat <int>, hdop <dbl>, vdop <dbl>,
#   pdop <dbl>, ageofdgpsdata <dbl>, dgpsid <int>,
#   ns3_TrackPointExtension <chr>, geometry <POINT [°]>

[datawookie]

Hi @ColinFay, wow! I was not aware that sf::read_sf() could consume GPX files. I think that this is awesome and would make process_data() a lot cleaner and easier to maintain. How about a PR? :) Thanks, Andrew.

[ColinFay]

Super, will do a PR !