Coerce an object to clusterTrack data format

S3 generic for converting objects into a ctdf.

Usage

as_ctdf(x, ...)

Arguments

x: An object to convert.
...: Passed to methods.

Value

A ctdf: a timestamp-ordered data.table with an sfc_POINT geometry column named location.

Details

See as_ctdf.data.frame() for coordinate columns and as_ctdf.sf() for existing POINT geometries with a source CRS.

Internal working columns used by the clustering pipeline are preallocated when a ctdf is created and reset after cluster_track() completes. Intermediate clustering states can be inspected with putative_cluster_trace() when trace = TRUE.

ctdf columns

In addition to columns retained from the input, a ctdf contains the following standardized columns:

timestamp: Observation time. Rows are ordered by this column when the ctdf is created.
location: Point geometry in the target coordinate reference system.
.id: Sequential row identifier assigned after ordering by timestamp. It is used internally to track locations through the clustering workflow and does not refer to the row number in the original input.
cluster: Final cluster assignment. It is initialized to NA. cluster_track() assigns positive integers to clustered locations and 0 to unassigned locations. aggregate_ctdf() may subsequently merge and renumber clusters.
lof: Local Outlier Factor score for each location. It is initialized to NA and populated by ctdf_lof() for locations with cluster > 0. Unassigned locations retain NA.