How to Create Network Graph Visualizations in Microsoft PowerBI

Microsoft PowerBI is a probably the most fashionable Business Intelligence (BI) instruments, and whereas it has all of the options it is advisable create dynamic analytic reporting for stakeholders throughout the enterprise, creating some superior knowledge visualizations is tougher.

This text will stroll by way of easy methods to create massive community graph visualizations in Microsoft PowerBI to allow dynamic and interactive exploration of interconnected datasets similar to provide chain networks, monetary transactions, and far more.

However earlier than we do this, let’s check out some fast foundations of community graphs. 

Community Graph Foundations

Information for community graphs, known as “graph knowledge” is knowledge formatted in node and edge format. Nodes symbolize discrete issues and edges symbolize the relationships between nodes.

Let’s take a easy instance of a web based social community, which could be represented in graph format. 

Nodes seek advice from profiles, whereas edges seek advice from following standing. 

A easy community of three profiles would possibly find yourself trying like this:

When visualizing community graphs, we will embed further details about nodes and edges in varied methods, similar to however not restricted to:

1. Node measurement
2. Edge measurement
3. Node colour
4. Edge colour
5. Labels

Structuring Community Information

So now that you realize the fundamental constructing blocks of a community graph, how do you construction and remodel your dataset?

Graph Information is In every single place

When you could be considering, “we solely have relational knowledge the place I’m at”, that’s typically not the case. Actually, plenty of relational datasets could be visualized as a community graph.

Let’s take a easy gross sales desk for instance with columns for product title, buyer title, and amount. 

We will symbolize this similar gross sales desk as a community graph by representing each product title because the node sort “product”, buyer title because the node sort “buyer”, and every row as the sting “Bought”.

Visualized as a community graph, this would possibly look one thing like:

Graph Information Codecs

There are a couple of methods this knowledge is structured, similar to however not restricted to:

1. Node & Edge Lists (Typically in .csv format)
2. Graph Databases (Akin to Neo4j)
3. Graph Information (similar to GraphML or GEXF)

However on this article, we will probably be utilizing a mixed node and edge record right into a single tabular dataset because of the necessities of creating community graphs inside Microsoft PowerBI. 

Mapping Your Information

You’ll have to map your knowledge to the next tabular format with every file representing an edge:

1. Supply Node (Required) -> It is a distinctive identifier of the beginning node of the sting (for instance, Buyer ID)
2. Goal Node (Required) -> It is a distinctive identifier of the ending node of the sting (for instance, Product ID)
3. Supply Coloration -> It is a class identifier for the supply node (for instance, Buyer Kind)
4. Goal Coloration -> It is a class identifier for the goal node (for instance, Product Class)
5. Hyperlink Coloration -> It is a class identifier for the sting (for instance, Gross sales Channel)

Creating the Community Graph Visualization

Now that we’ve our knowledge mapped, we will create the community graph visualization. 

Whereas Microsoft doesn’t embody a community visible within the default PowerBI visuals, we will entry the visible market to obtain third-party visuals.

For this text, we will probably be utilizing the visible “Astra”, which helps you to create large-scale community graphs with loads of customization choices.

Upon getting it put in, it will likely be in your visible library.

Drag the visible onto your canvas, choose it, and word the values required (which we mapped earlier). The visible additionally has choices to cross x and y coordinates in addition to customized labels, nonetheless we received’t use these choices on this article.

The one required values are “Supply Node” and “Goal Node” so let’s begin there. Drag the columns you mapped to these nodes from the info pane.

You’ll discover the visible graphs our nodes and edges, nonetheless, it isn’t trying so nice. We’ll want to vary a few of the simulation settings.

To vary the simulation settings, open the formatting pane, then simulation, and improve each the hyperlink distance and repulsion power. I selected to set repulsion to 0.3, and hyperlink distance to fifteen.

Now you can see that we get a a lot better structure of our knowledge.

Let’s now encode some further data into the graph, by altering the node colour primarily based on node classes. Drag the fields you mapped above to Supply Coloration and Goal Coloration.

You’ll now discover the nodes are coloured otherwise and we’ve a legend on the visible. 

 Let’s do some formatting to the background colour and node colours within the formatting pane.

Congratulations! You’ve created a community graph visualization in PowerBI with dynamic node coloring. 

We add much more data to the graph, for instance:

• Activate node weight to make nodes with extra edges bigger in measurement
• Including a hyperlink class to the colour the hyperlinks
• Including totally different labels to the nodes

However we aren’t completed there.

As soon as we’ve the visualization, stakeholders have to make use of it to make extra knowledgeable choices.

Interacting with the Community Graph

There’s quick worth in a static community graph, similar to having the ability to visually see how knowledge is interconnected by way of relationships. 

Nonetheless, there are some further options we will use to make the visualization extra insightful. 

First, we will work together with the legend by deciding on classes to focus on them on the graph. For instance, shortly finding Widgets within the graph:

We will additionally choose particular person nodes within the graph by clicking on them.

Alternatively, you possibly can toggle “choose adjoining nodes” within the node properties to have it choose not simply the node clicked on, however all nodes straight linked to it by way of an edge.

For instance, deciding on “Widget A” with “choose adjoining nodes” on exhibits all clients who’ve bought that widget:

However deciding on nodes doesn’t simply spotlight them within the visualization, it passes that filter to the remainder of your PowerBI report.

This implies we will add further charts to present some extra context to the person’s picks. 

For instance, including a bar chart for amount bought by buyer:

We will additionally do the reverse by filtering the info going into the community visible. This may be completed in a number of methods, similar to:

1. Slicers
2. Choosing items of different charts, similar to a slice of a donut chart
3. Filter pane

Let’s use a slicer to slice the graph on Buyer Kind:

Constructing Advanced BI Studies

Whereas the instance community graph on this article is comparatively easy for demonstration functions, you possibly can construct fairly advanced BI reporting for stakeholders. 

The Astra PowerBI visible used on this article can scale to tons of of hundreds of edges, and paired with further cross-filtered visuals & slicers can allow extra superior analytics than is feasible with default PowerBI stories.

Conclusion

Community graphs are throughout us, even hiding in your relational datasets. Whereas there’s nice community graphing tooling on the market, constructing community graphs in PowerBI lets you convey this superior analytic instrument to your commonplace BI stakeholders, in addition to construct superior reporting by including context with further filters and charts.