Search This Blog

Saturday, February 10, 2018

Power BI with different Network Visualizations

(2018-Feb-10) A few days ago Microsoft provided another update for its desktop version of the Power BI data analytical tool - https://powerbi.microsoft.com/en-us/blog/power-bi-desktop-february-2018-feature-summary/. Among all the major updates in this release, there was one new custom network visualization added by ZoomCharts. It wasn't the first network visualization available in the Power BI and I played a little bit with some of them in the past.

Currently, there are 5 network visual controls that you can add to your Power BI reports:
- Network Navigator Chart by Microsoft
- Force-Directed Graph by Microsoft
- Journey Chart by MAQ Software
- Social Network Graph by Arthur Graus
- Network Visualization by ZoomCharts

I used an open dataset from one of my previous reports based on the City of Toronto Lobbyist Registry and for testing various network visualizations I've used a portion of this data that describes all the communications between registered lobbyist and city officials.





Here is a data model of the original Power BI report and for this test case I only used the highlighted la_communication table:


Most of the network visualizations in Power BI share a similar concept, your dataset needs to have a least one descriptive attribute for a Source and another one as a Target for your network relationship. Let's take a look how all those five visual components are similar and different to each other.



Network Navigator Chart
https://appsource.microsoft.com/en-us/product/power-bi-visuals/WA104380795?&tab=Overview












Network Navigator lets you explore node-link data by panning over and zooming into a force-directed node layout (which can be precomputed or animated live).

From an initial overview of all nodes, you can use simple text search to enlarge matching nodes in ways that guide subsequent navigation. Network nodes can also be color-coded based on additional attributes of the dataset and filtered by linked visuals.

You can set source and target nodes using your dataset, it would also help if you add additional attributes to set different colors for your nodes. The difficult part is that with a very large dataset, this visualization becomes very slow to process and animate all the relationships between the nodes.



Force-Directed Graph
https://appsource.microsoft.com/en-us/product/power-bi-visuals/WA104380764?&tab=Overview













This custom visual implements a D3 force layout diagram with curved paths. The thickness of the path represents the weight of the relationship between the nodes.


Since the relationship and interconnection between a large set of entities could be very complex, the visual positions the nodes in such a way that there are few crossings as possible, making the exploration experience easy, fun. Users can also adjust the layout manually by simply dragging the nodes around.

This visualization is my favorite, it has its own animation engine. Relationships between a source and target nodes are marked with arrows which help to identify who initiates a communication and who is a recipient.


Journey Chart
https://appsource.microsoft.com/en-us/product/power-bi-visuals/WA104380989?&tab=Overview











Journey Chart by MAQ Software enables users to display complex, multi-stage lead paths in a clean, uncluttered visual. Nodes represent categories and vertices represent relationships between categories. The bigger the node or vertex, the larger the value. Customizable node colors make distinguishing between categories easy.

This visualization doesn't have a source or target nodes, it rather creates a metric journey between several categories (two minimum); it also has a root node and then it shows the first level of descriptive category and a relationship is then built using next level category. It's fun to play with this chart, you can move nodes and unfold its vertices; I would highly recommend this visual chart for your own data discovery.


Social Network Graph
https://appsource.microsoft.com/en-us/product/power-bi-visuals/WA104381236?&tab=Overview













Super easy network visualization to use; it has standard source and target nodes that could be linked to you dataset attributes. Along with basic node relationships, you can assign graphics node representations (aka photos) using URL based links. However, if you don't have those links then your report's network graph looks like a bunch of empty animated circles because the current version of this visualization doesn't support text label settings, only tooltip messages are available.


Network Visualization
https://appsource.microsoft.com/en-us/product/power-bi-visuals/WA104381419?tab=Overview











ZoomCharts Network Visual (https://zoomcharts.com/en/microsoft-power-bi-custom-visuals/) enables tabular data visualization, exploration, and filtering using network layout. Clicking on any node acts as a filter for the rest of the dashboard, making it easy to drill-through data and boosting Power BI productivity.

In one way this network chart resembles the Journey chart by MAQ Software since it allows you to list all of the chart nodes by listing them in the order, like categories. From the other angle, it works like two other Microsoft mainstream network visualizations: you can see how nodes relationships are constructed and metric values define the size of each node. Plus the ZoomCharts network visualization has additional user controls that you can adjust right within your Power BI report, which is super cool! The one thing that I didn't like about this chart was that I couldn't hide or remove legend of my nodes (Lobbyist and POH Office name in my dataset case).

And, if you have reached to this point of reading this blog post, here is the actual Power BI report, fully interactive and accessible. I will let you decide which network Power BI visualizations you would like to use, and I already have my own preferences :-)




(2018-Feb-11) Update: On the set of unfiltered 13177 records of Source to Target pairs, the fastest network visualization in Power BI became the visual control by ZoomCharts and the slowest one was Force-Directed Graph by Microsoft.

Here is a brief summary of testing each of the individual custom visualizations. 

- Network Navigator - 20 seconds
- Force-Directed Graph - more than 300 seconds and I still didn't get the final network finished rendering (it works well with smaller datasets though).
- Journey Chart by MAQ Software - 120 seconds
- Social Network Graph - 40 seconds
- Network Visualization - 15 seconds

5 comments:

  1. Nice work mate! Any advice on which ones provide the best performance?

    ReplyDelete
  2. Thanks Simon, I've done some additional testing and updated the post:
    On the set of unfiltered 13177 records of Source to Target pairs, the fastest network visualization in Power BI became the visual control by ZoomCharts and the slowest one was Force-Directed Graph by Microsoft.

    ReplyDelete
  3. It's so nice article thank you for sharing a valuable content. Power BI Online Training

    ReplyDelete
  4. Thanks for sharing this post. Your post is really very helpful its students. Power BI Online Training

    ReplyDelete