TRK_FindDwells

TRK_FindDwells takes a track, a distance threshold, and a duration threshold and returns an array of tracks. The result is all the dwells detected from the input track based on the distance threshold and duration threshold provided.

Distance threshold is the maximum distance between point observations in a track to be considered a single dwell. Duration threshold is the minimum duration of a track to be considered a single dwell.

Dwell is defined as a track that contains sequential observations with little or no movement over a certain period of time.

Tracks are linestrings that represent the change in an entity's location over time. Each vertex in the linestring has a timestamp (stored as the M-value) and the vertices are ordered sequentially.

For more information on using tracks in GeoAnalytics Engine, see the core concept topic on tracks.

Function	Syntax
Python	`find_dwells(track, distance_threshold, duration_threshold)`
SQL	`TRK_FindDwells(track, distance_threshold, duration_threshold)`
Scala	`findDwells(track, distanceThreshold, durationThreshold)`

For more details, go to the GeoAnalytics Engine API reference for find_dwells.

Python and SQL Examples

Python

SQL

Use dark colors for code blocksCopy
from geoanalytics.tracks import functions as TRK
from geoanalytics.sql import functions as ST
from pyspark.sql import functions as F

data = [(1, "POINT (0 0)", "2021-10-05 10:30:00"),
        (1, "POINT (5 0)", "2021-10-05 10:31:00"),
        (1, "POINT (10 0)", "2021-10-05 10:32:00"),
        (1, "POINT (11.2 1.1)", "2021-10-05 10:33:00"),
        (1, "POINT (12.2 -0.2)", "2021-10-05 10:34:00"),
        (1, "POINT (11.2 -1.2)", "2021-10-05 10:35:00"),
        (1, "POINT (10.5 -0.9)", "2021-10-05 10:36:00"),
        (1, "POINT (10.6 0.5)", "2021-10-05 10:37:00"),
        (1, "POINT (11.5 0)", "2021-10-05 10:38:00"),
        (1, "POINT (11.6 -0.5)", "2021-10-05 10:39:00"),
        (1, "POINT (10.5 -0.3)", "2021-10-05 10:40:00"),
        (1, "POINT (12 0)", "2021-10-05 10:41:00"),
        (1, "POINT (12.5 0)", "2021-10-05 10:42:00"),
        (1, "POINT (18 0)", "2021-10-05 10:43:00"),
        (1, "POINT (20 0)", "2021-10-05 10:44:00")]

df = spark.createDataFrame(data, ["id", "wkt", "datetime_str"]) \
    .withColumn("point", ST.point_from_text("wkt", 3857)) \
    .withColumn("timestamp", F.to_timestamp("datetime_str")) \
    .drop("wkt", "datetime_str")

df_trk = df.groupBy("id").agg(TRK.aggr_create_track("point", "timestamp").alias("track"))

trk_dwells = df_trk.select(TRK.find_dwells("track", ST.create_distance(3, 'Meters'), ST.create_duration(1, 'Minutes')).alias("dwells"))

axes = df_trk.st.plot("track", edgecolor="lightgrey", linewidths=10, figsize=(15, 8))
trk_dwells.select(F.explode("dwells")).st.plot(ax=axes, edgecolor="greenyellow", linewidths=3)
axes.legend(['Input track','Result track'], loc='lower right', fontsize='large', bbox_to_anchor=(1, 0))

Plotted result for TRK_FindDwells.

Scala Example

Scala
Use dark colors for code blocksCopy
import com.esri.geoanalytics.sql.{functions => ST}
import com.esri.geoanalytics.sql.{trackFunctions => TRK}
import org.apache.spark.sql.{functions => F}
import java.sql.Timestamp

case class pointRow(id: Int, pointWkt: String, timestamp: Timestamp)

val data = Seq(pointRow(1, "POINT (0 0)", Timestamp.valueOf("2021-10-05 10:30:00")),
               pointRow(1, "POINT (5 0)", Timestamp.valueOf("2021-10-05 10:31:00")),
               pointRow(1, "POINT (10 0)", Timestamp.valueOf("2021-10-05 10:32:00")),
               pointRow(1, "POINT (11.2 1.1)", Timestamp.valueOf("2021-10-05 10:33:00")),
               pointRow(1, "POINT (12.2 -0.2)", Timestamp.valueOf("2021-10-05 10:34:00")),
               pointRow(1, "POINT (11.2 -1.2)", Timestamp.valueOf("2021-10-05 10:35:00")),
               pointRow(1, "POINT (10.5 -0.9)", Timestamp.valueOf("2021-10-05 10:36:00")),
               pointRow(1, "POINT (10.6 0.5)", Timestamp.valueOf("2021-10-05 10:37:00")),
               pointRow(1, "POINT (11.5 0)", Timestamp.valueOf("2021-10-05 10:38:00")),
               pointRow(1, "POINT (11.6 -0.5)", Timestamp.valueOf("2021-10-05 10:39:00")),
               pointRow(1, "POINT (10.5 -0.3)", Timestamp.valueOf("2021-10-05 10:40:00")),
               pointRow(1, "POINT (12 0)", Timestamp.valueOf("2021-10-05 10:41:00")),
               pointRow(1, "POINT (12.5 0)", Timestamp.valueOf("2021-10-05 10:42:00")),
               pointRow(1, "POINT (18 0)", Timestamp.valueOf("2021-10-05 10:43:00")),
               pointRow(1, "POINT (20 0)", Timestamp.valueOf("2021-10-05 10:44:00")))

val points = spark.createDataFrame(data)
            .withColumn("point", ST.pointFromText($"pointWkt", F.lit(3857)))

val tracks = points.groupBy("id").agg(TRK.aggrCreateTrack($"point", $"timestamp").alias("track"))

val dwellTracks = tracks.select(TRK.findDwells($"track", ST.createDistance(3, "Meters"), ST.createDuration(1, "minutes")).alias("dwells"))

dwellTracks.show(truncate = false)

Result
Use dark colors for code blocksCopy
+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|dwells                                                                                                                                                                                                                     |
+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
|[{"hasM":true,"paths":[[[10,0,1.63345512e9],[11.2,1.1,1.63345518e9],[12.2,-0.2,1.63345524e9],[11.2,-1.2,1.6334553e9],[10.5,-0.9,1.63345536e9],[10.6,0.5,1.63345542e9],[11.5,0,1.63345548e9],[11.6,-0.5... (280 characters)]|
+---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+

Version table

Release	Notes
1.5.0	Python, SQL, and Scala functions introduced

TRK_FindDwells

Python and SQL Examples

Scala Example

Version table

Related topics