Find Closest Facilities finds the given number of facilities that are reachable from each incident within the specified travel time or travel distance, and returns the best routes between the incidents and the chosen facilities. When finding closest facilities, you can specify whether the direction of travel is toward or away from the facilities. For example, the tool can be used to find the closest fire stations to fire incidents, the closest healthcare providers to a person's address, or the closest hospitals for emergency response.
Usage notes
-
Find Closest Facilities requires two point DataFrames representing the incidents and the facilities.
-
A network dataset is required to run any network analysis tool. It must be locally accessible to all nodes in your Spark cluster. Use
set
to load the network dataset from a mobile map package or a mobile geodatabase.Network() -
A travel mode refers to the mode of transportation, such as driving or walking. Use
set
to choose a mode defined in the network datasource, or a custom travel mode defined in a JSON string. By default, the tool uses the default travel mode in the network datasource.Travel Mode() -
Use the
set
setter to specify the direction of travel to or from the facilities.Travel Direction() -
T
—The closest facilities are searched along the network from the incidents to the facilities within the specified impedance cutoff. This is the default.o Facilities -
From
—The closest facilities are searched along the network from the facilities to incidents within the specified impedance cutoff.Facilities
The travel direction you should use depends on your use case. For example, in emergency response,
From
is commonly used to calculate the travel time or travel distance from the facilities (e.g., hospitals, fire stations, police stations) to the locations of the emergency. In retail store management,Facilities T
is commonly used to calculate the travel time or travel distance from customers' addresses to the retail stores to analyze the proximity of the facilities.o Facilities -
-
It is required to set the impedance cutoff using
set
. It accepts a single cutoff value. The impedance cutoff is used to set the maximum travel distance or travel time when searching for facilities for each incident. The impedance cutoff must be the same units as the travel mode. For example, if the travel mode is driving time, the impedance cutoff should be time based.Cutoff() There are two types of cutoff value supported in the Find Closest Facilities tool.
-
Distance cutoff—Specify the maximum traveling distance between incidents and facilities. For example, when analyzing walking distance from schools (incident DataFrame) to subway stations (facility DataFrame), a cutoff value of 1 mile (e.g.,
set
) means that the tool will search for the closest subway stations within 1 mile walking from each school.Cutoff(1, "mile") -
Time cutoff—Specify the maximum traveling time between incidents and facilities. For example, when analyzing driving time from fire stations (facility DataFrame) to fire incidents (incident DataFrame), a cutoff value of 15 minutes (e.g.
set
) means the tool will search for the closest fire stations within a 15-minute drive to the fire incidents.Cutoffs(15, "minutes")
The cutoff value must be positive. If the unit is missing, the tool will use the distance or time unit defined in the travel mode.
-
-
Use the
set
setter to specify the maximum number of closest facilities to find per incident. If there are multiple facilities with an equal travel cost to an incident, Find Closest Facilities will break ties by randomly selecting one or more records from the equidistant facilities to ensure the specified number of closest facilities. For example, if you are interested in finding two closest facilities when there are three facilities that are equidistant from the incident, two of the three facility records will be randomly selected and returned in the output.N u m Facilities() -
The impedance cutoff can result in fewer facilities returned than the specified number of facilities. For example, if you are interested in finding three closest facilities within a specified travel distance when there are two identified facilities within the impedance cutoff, only the two facilities will be returned in the output. When there are no closest facilities found within a specified cutoff, the tool will return a
Null
value in the output for no facilities. -
Find Closest Facilities can return the best routes or straight lines between the incidents and the chosen facilities. You can also choose not to return line geometry for better performance. Use
set
to set the route geometry for the output DataFrame:Route Geometry() -
Along
—The true shape of the resulting route that is based on the streets along the network.Network -
Straight
—A straight line between the incident and the identified facility.Lines -
N
—No line geometry will be returned.o Lines
When the output route geometry is set to
Along
orNetwork Straight
, the primary geometry field of the output Dataframe is the route geometry field. When set toLines Nolines
, there is no primary geometry field in the output Dataframe. -
-
Use the setter
accumulate
to specify cost attributes accumulated along the network. The cost attributes are defined in the network dataset. One or moreAttributes() Total_
columns will be returned, where[Cost] Cost
is the name of the cost attribute. For example, if the available cost attributes in the network dataset areKilometers
,Minutes
, andWalk
, you can accumulate all attributes by callingTime accumulate
. In this case three output fields (Attributes("Kilometers", "Minutes", "Walk Time") Total_
,Kilometers Total_
, andMinutes Total_
) will be returned, representing the cost along the network between the associated origin and destination.Walk Time -
The analysis will always be completed in the coordinate system of the network dataset. If the incident Dataframe or facility Dataframe are in a different coordinate system than the network dataset , they will automatically be transformed to the coordinate system of the network dataset.
Limitations
-
Network analysis requires a network dataset from a mobile map package or a mobile geodatabase. Loading network data from a file geodatabase is not supported. Using a network service, such as the ArcGIS Online network analysis service, is not supported.
-
GeoAnalytics Engine does not support adding barriers in network analysis.
-
GeoAnalytics Engine does not support using traffic info in network analysis.
Results
The following fields are included in the output DataFrame:
- All fields from the incident DataFrame
- All fields from the facility DataFrame
In addition, the following fields representing the travel cost are included for each output record:
Field | Description |
---|---|
Rank | The rank of the closest facilities. The rank is given according to ascending-order travel distance or time. |
Travel | The travel time in minutes between the incident and the identified facility. |
Travel | The travel distance in meters between the incident and the identified facility. |
Both travel time and travel distance are calculated along the network. If you accumulate cost attributes,
your output will have one or more fields named Total_
representing the accumulated travel cost along the network
between the incident and identified facility.
If you set the output route geometry to Along
or Straight
, your output will have a linestring field named
route_
representing the true shape or straight line between the incident and the identified facility.
Performance notes
Improve the performance of the Find Closest Facilities tool by doing one or more of the following:
-
Only analyze the records in your area of interest. You can pick the records of interest by using one of the following SQL functions:
- ST_Intersection—Clip to an area of interest represented by a polygon. This will modify your input records.
- ST_BboxIntersects—Select records that intersect an envelope.
- ST_EnvIntersects—Select records having an evelope that intersects the envelope of another geometry.
- ST_Intersects—Select records that intersect another dataset or area of intersect represented by a polygon.
- Set the route geometry to
Nolines
instead of theAlong
orNetwork Straight
if you are only interested in determining the total travel time or travel distance between the incidents and facilities.Lines - Use smaller values for
set
andCutoff() set
.Number Facilities()
Similar capabilities
Syntax
For more details, go to the GeoAnalytics Engine API reference for find closest facilities.
Setter | Description | Required |
---|---|---|
run(incidents_ | Runs the Find Closest Facilities tool using the provided DataFrames. | Yes |
set | Sets the network data source from a mobile map package or a mobile geodatabase. | Yes |
set | Sets the travel mode. By default, the tool uses the default travel mode in the network datasource. | No |
set | Sets the direction of travel to or from the facilities. Choose from 'To (default) or 'From . | No |
set | Sets the maximum travel distance or travel time searching for facilities for each incident. By default, it is in the unit of the impedance attribute used by the travel mode. | Yes |
set | Sets the number of facilities to find. The default is 1. | No |
set | Sets the route geometry representing the route between the incident and the identified facility. Choose from 'Along (default), 'Straight , or 'No . | No |
accumulate | Accumulates cost attributes along the network between the incident and the identified facility. No accumulated cost is returned by default. | No |
Examples
Run Find Closest Facilities
# Log in
import geoanalytics
geoanalytics.auth(username="myusername", password="mypassword")
# Imports
from geoanalytics.tools import FindClosestFacilities
from pyspark.sql import functions as F
import matplotlib.pyplot as plt
# Create a facility DataFrame
facilities_url = r"https://services.arcgis.com/P3ePLMYs2RVChkJx/ArcGIS/rest/services/SDFireStations/FeatureServer/0"
facilities_df = spark.read.format("feature-service").load(facilities_url) \
.select("FACILITYID", "FULLADDR", "PHONE", "ACTIVE", "shape")
# Create an incident DataFrame
incidents_url = r"https://services1.arcgis.com/Ua5sjt3LWTPigjyD/arcgis/rest/services/Public_School_Location_201819/FeatureServer/0"
incidents_df = spark.read.format("feature-service").load(incidents_url) \
.filter(F.col("NMCNTY") == 'San Diego County') \
.select("NAME", "STREET", "CITY", "STATE", "ZIP", "shape")
# Access the Network Dataset
# This needs to be accessible to the machine that is running the Find Closest Facilities tool.
# If running on a cluster, it needs to be accessible to all nodes in the cluster.
california_network = r"/data/California.mmpk"
# Run the Find Closest Facilities tool
result = FindClosestFacilities() \
.setNetwork(california_network) \
.setTravelMode("trucking time") \
.setTravelDirection("FromFacilities") \
.setCutoff(5, "minutes") \
.setNumFacilities(2) \
.setRouteGeometry("AlongNetwork") \
.run(incidents_df, facilities_df) \
.where(F.col("TravelTime").isNotNull())
result.sort("NAME","Rank").show(5)
+--------------------+---------------+---------+-----+-----+--------------------+----+------------------+------------------+----------+--------------------+------------+------+--------------------+--------------------+
| NAME| STREET| CITY|STATE| ZIP| shape|Rank| TravelTime| TravelDistance|FACILITYID| FULLADDR| PHONE|ACTIVE| shape1| route_geometry|
+--------------------+---------------+---------+-----+-----+--------------------+----+------------------+------------------+----------+--------------------+------------+------+--------------------+--------------------+
| ALBA|4041 Oregon St.|San Diego| CA|92104|{"x":-117.1348499...| 1|2.7688534783676917|1002.3922083982721| 14| 4011 32nd Street|619-533-4300| Yes|{"x":-117.1249271...|{"paths":[[[-117....|
| Adams Elementary| 4672 35th St.|San Diego| CA|92116|{"x":-117.1180409...| 1|1.7452089199273755| 695.6191880289591| 18| 4676 Felton Street|619-533-4300| Yes|{"x":-117.1220022...|{"paths":[[[-117....|
| Adams Elementary| 4672 35th St.|San Diego| CA|92116|{"x":-117.1180409...| 2| 4.719640574786083| 1959.747748980926| 14| 4011 32nd Street|619-533-4300| Yes|{"x":-117.1249271...|{"paths":[[[-117....|
|Albert Einstein A...| 3035 Ash St.|San Diego| CA|92102|{"x":-117.1384500...| 1| 2.156189057040781| 737.2832135729769| 11| 945 25th Street|619-533-4300| Yes|{"x":-117.1400096...|{"paths":[[[-117....|
|Albert Einstein A...| 3035 Ash St.|San Diego| CA|92102|{"x":-117.1384500...| 2| 4.247430230528537| 1439.274596232985| 7|944 Cesar E. Chav...|619-533-4300| Yes|{"x":-117.1450304...|{"paths":[[[-117....|
+--------------------+---------------+---------+-----+-----+--------------------+----+------------------+------------------+----------+--------------------+------------+------+--------------------+--------------------+
only showing top 5 rows
Plot results
# Plot the closest fire stations near public schools in San Diego
colors = ['#e34a33', '#fdcc8a']
cmap = plt.cm.colors.ListedColormap(colors)
# Plot the true routes
result_plot = result.st.plot(cmap_values="Rank",
cmap=cmap, is_categorical=True,
basemap="light",
figsize=(15,15),
legend=True,
legend_kwds = {"title": "Rank of closest fire stations"})
# Plot the public school in green
result.st.plot(geometry="shape", facecolor = "#1a9641", marker_size=30, ax=result_plot)
# Plot the closest fire stations near public schools
result.st.plot(geometry="shape1",
cmap_values="Rank",
cmap=cmap, is_categorical=True,
marker_size=30,
ax=result_plot)
result_plot.set_title("Closest fire stations and routes to public schools in San Diego")
result_plot.set_xlabel("X (Degrees)")
result_plot.set_ylabel("Y (Degrees)")
Version table
Release | Notes |
---|---|
1.3.0 | Tool introduced |