Generic selectors
Exact matches only
Search in title
Search in content
Post Type Selectors
podcast
Filter by Categories
Galleries
Print Category 1
Print Category 2
Uncategorized

Working with Networks in GIS

Between the Lines: Working with Networks in GIS

Figure 1: An electrical utility network with coinciding diagram using top-down orientation.

Geospatial information systems, otherwise known as GIS, is gaining popularity in industries across the world due to its ability to provide analytical and business insights using platform- and service-based systems. One large contributor to this rise in popularity is the ability to manage assets and gauge their effectiveness using network datasets. 

The three most common network dataset types used in the ArcGIS ecosystem

Utility networks

Trace network

Transportation networks (or simply Network Analyst networks) 

Electrical, telecommunications, water, and gas companies deploy utility networks to manage their assets and isolate impact areas when problems occur. Conversely, companies involved in logistics or transportation use network analyst datasets for routing and analysis within road networks.

Newer, but just as impactful, are trace networks that are utilized by rail and hydrology companies to perform analysis to trace flow direction within their network.

In this article, we’ll discuss what the different types of networks have in common, then provide an overview of unique functionality and real-world applications for each.

Shared Characteristics of Spatial Network Datasets

Each of the three types of networks mentioned above have similar characteristics

At the most basic level, the networks are comprised of simple features – points and lines. 

When the network is created, the network stores connectivity between the source features using a series of connected edges and junctions.

For instance, in a road network dataset created with the network analyst extension, a stoplight intersection is stored as road line features (edges) and point features where they intersect (junctions). Without the network, these features would be unaware of each other, and analysis would be unable to be performed as the features would lack the context they need to route through each other.

It is important to note that all features in these networks must be in the same feature dataset (and coordinate system!). Feature datasets group related features together and can store network topology. 

Network topology is an advanced feature that makes data interactions in the network possible, and regulates what is acceptable behavior given the inherent rules of each network type.

For information about feature datasets, see the feature datasets in ArcGIS Pro documentation.

Tracing is one relevant advanced feature that is available with an established network. Tracing is possible through geoprocessing tools located within the geoprocessing toolbox for each type of networks. Tracing functionality will perform analysis and return existing connectivity based on the connectivity and traversability of input features used in the tool. Typically, this will involve starting (origin) points and may include end (destination) points. One can also input barriers that prevent flow through a certain area of the network.

Figure 2: Geoprocessing toolboxes with tracing functionality pointed out.

As edits are made to an established network, utility, or trace network, “dirty areas” will appear on the map. These will show as hatched polygons within the bounding box of the inserted features.

Figure 3: An example of “dirty areas” when inserting edits into a utility network.

Dirty areas must be validated in the network topology before edits are saved. This ensures the geometric coincidence (the x, y, and z coordinates) between the related features stays intact as features are added or changed. This prevents you from accidentally ‘breaking’ your network. 

Additionally, the various types of networks have data models that help them function. The data model can be thought of as the framework, or template, that the network operates under. In a utility network, this would be the schema (water, gas, electrical, etc.) that the network features exist in.

In a road network dataset, this would be different modes of transportation (i.e., passenger vehicle driving vs. pedestrian walking) that store characteristics for analysis within that mode of transportation.

Lastly, each network is able to function as a single-user deployment within a file geodatabase or as a multi-user deployment using enterprise geodatabases, and traditional or branch versioning.

Network Analyst Functionality and Applications

Figure 4: A map displaying an Origin-Destination Cost Matrix – One of the analysis methods available using Network Analyst datasets.

Network analyst datasets are frequently used to create road networks for Department of Transportation or logistics companies.

Analysis within network analyst networks is made possible through transportation modes and associated characteristics. The most common characteristics are costs, restrictions, and descriptor attributes which are unique to Network Analyst networks.

Costs, also known as impedance, are a measurement that helps find the best path, or route, in the analysis. The values of a cost measurement are dependent on the length of the line (or edge) element, otherwise known as an apportioned measurement. Costs are generally presented as time or distance. For example, the same trip could be described as having a cost of 15 minutes, or a cost of 4 miles. 

Restrictions are true/false values that set “prohibited” criteria. True/false values are also known as Boolean type values. As such, they are either a yes or a no. If a restriction is violated, that part of the route in no longer considered in the analysis. 

An example of this would be a walking restriction preventing pedestrians from walking on highways. Roads in the network analyst road network defined as highways would be excluded from the analysis as it would be dangerous for pedestrians to walk on them.

Restrictions can be applied to different types of transportation modes to customize analysis and better mimic real-world conditions.

Descriptor attributes provide additional information that can help define hierarchical rules to supplement restrictions in the network dataset. Think of hierarchical rules being applied in situations where multiple criteria are desired in the analysis, but one criterion is more important than the other. Descriptors allow this prioritization to be possible.

Typical analyses with network analyst road networks are speed limits, or bridge clearance heights. This additional information about the roads can be used to add richer detail to the analysis, and these attributes may be leveraged in calculating costs or applying restrictions. 

Figure 5: An infographic representing different path calculations using different modes of transportation and cost (impedance) measurements – a) Automobile time calculation, b) Automobile distance calculation, c) Walking distance calculation.

Utility Network Functionality and Applications

Utility Networks provide applications for managing assets within the electric, gas, water, stormwater, wastewater, and telecommunications industries. 

Note: Utility networks are an immense topic, and successfully deploying a utility network is challenging. Providing a thorough description of all the factors involved in successfully deploying a utility network is beyond the scope of this article as the article is only intended to highlight utility network functionality and applications. For further explanation, please consult the resources linked in this section.

Building on concepts discussed in the network analyst section, utility networks provide connectivity associations between point and line features participating in the utility network. Utility networks operate within specific schemas to manage data within the utility network. Esri provides ready-made Asset Packages for use with utility networks, but customization is possible.

Typical schemas from this Esri blog are shown below:

Graphical user interface

Description automatically generated with medium confidence
Figure 6: Example utility network schemas

What makes utility networks special is their ability to create networks and subnetworks based on these connectivity associations.

Subnetworks can be fine-tuned down to an individual valve in water distribution datasets, or individual fuses in electrical datasets. You can almost think of them as a network within a network. 

This provides an organization with an additional item in their toolkit to manage and model their assets, perform geospatial analysis on them, and troubleshoot solutions to outages or leaks within the network. 

As an additional feature to assist in understanding the complexities of their systems, organizations can build network diagrams to visualize their networks and subnetworks within the utility network. Activated by using trace functionality, networks diagrams are essentially “schematics” that provide a symbolic representation of the associated features in the utility network. These are especially useful for describing functionality to those without a GIS background.

Figure 7: A network diagram representing an electrical utility network.

The impacts of utility network analyses are made doubly useful by housing several interrelated features within a single polygon feature, called a container.

 Containment features can track internal paths within an object, such as an electrical transformer, and are the crux that allow the fine-tuning for managing flow, isolating issues, and creating subnetworks as mentioned above. For a more detailed explanation, see the containment associations documentation.

Trace Network Functionality and Applications

Trace networks are designed specifically for non-utility users who still need network tracing capabilities. 

This more flexible data model provided is a newer version of legacy geometric network functionality and is currently intended for hydrology, sewer, and railway applications. 

Trace networks have limited functionality when compared to utility networks, but provide a more streamlined approach for applications that do not need the intricacies and implementation cost associated with utility networks. Learning to work with a trace network is a great stepping stone for working with network analyst datasets, or the utility network. 

While one could take a crack at visualizing the limited functionality of trace networks compared to utility networks, the chart provided by Esri in the image below captures it splendidly:

Table

Description automatically generated
Figure 8:Image from ESRI’s Introducing the Trace Network blog.

An example of a trace network application using data from the National Hydrography Dataset (NHD) is shown in the image below. National Hydrography Dataset download instructions are available here.

Figure 9:  Visualization of two different types of traces performed on a trace network representing a stream network in the state of Maine, United States: a) Upstream trace and b) Downstream trace. The green dot in each image is the junction point used in the trace. In a), the highlighted area shows all streams that flow into (upstream) the junction. In b), the highlighted area shows all stream that flow out of (downstream) the junction.