Appends parsed Twitter data with latitude and longitude variables using all available geolocation information.
Usage
lat_lng(
x,
coords = c("coords_coords", "bbox_coords", "geo_coords"),
prefs = "bbox_coords"
)
Arguments
- x
Parsed Twitter data as returned by various rtweet functions. This should be a data frame with variables such as "bbox_coords", "coords_coords", and "geo_coords" (among other non-geolocation Twitter variables).
- coords
Names of variables containing latitude and longitude coordinates. Priority is given to bounding box coordinates (each obs consists of eight entries) followed by the supplied order of variable names. Defaults to "bbox_coords", "coords_coords", and "geo_coords") (which are the default column names of data returned by most status-oriented rtweet functions).
- prefs
Preference of coordinates to use as default, must be in
coords
.
Details
On occasion values may appear to be outliers given a previously used query filter (e.g., when searching for tweets sent from the continental US). This is typically because those tweets returned a large bounding box that overlapped with the area of interest. This function converts boxes into their geographical midpoints, which works well in the vast majority of cases, but sometimes includes an otherwise puzzling result.
See also
Other geo:
lookup_coords()
Examples
if (auth_has_default()) {
## stream tweets sent from the US
rt <- search_tweets(geocode = lookup_coords("usa"))
## use lat_lng to recover full information geolocation data
rtl_loc <- lat_lng(rt)
rtl_loc
}