Each of the tyGraph Models combine a concise perspective of your organization in an efficient and standardized model. However, sometimes you have another data source you want to combine atop the model we already have.
In this article we will show how to connect existing tyGraph Power BI Datasets to other data sources.
TABLE OF CONTENTS
- Step 1: Prepare Power BI Desktop
- Step 2: Choose Your Start File
- Step 3: Connect First Dataset
- Step 4: Add Additional Data Source
- Step 5: Modeling and Publication
- Summary
Step 1: Prepare Power BI Desktop
You will need a few things before we start:
A) You will need to start in Power BI Desktop and be running at least the Dec 2020 update.
B) Enable "Direct Query for Power BI Datasets and Analysis Services".
This feature is now Generally Available and this step can be skipped. To find out more, please refer to this blog post: Announcing general availability for composite models on Power BI Datasets and Analysis Services models | Microsoft Power BI Blog | Microsoft Power BI
C) Finally, sign in with your M365 login which is critical for connecting to artifacts in the Power BI service.
What is the difference between Power BI Desktop and Service?
Step 2: Choose Your Start File
You can start with a blank Power BI desktop model if you are looking to re-create report layouts from scratch. For those looking to add to default product, we recommend you download a live connected template for the product with layouts that you would like to use. This saves you re-creating every visual, on each canvas and gives you something to work from.
For example if I'm adding an extra user table for a tyGraph Pages dashboard then I would download the template for that tool so I can add my extra component but still have the default look and feel.
Template Files
| tyGraph for Teams | tyGraph for Yammer |
| tyGraph for SharePoint | tyGraph for OneDrive |
| tyGraph Pages | tyGraph Pulse |
Step 3: Connect First Dataset
If you're using a template file from above, open it from your downloads. If you're starting from scratch press get data>Power BI datasets and skip to step 4.
- Sign in as yourself in Power BI if you are not already. Power BI may prompt you with the following "enter your email address"
- You will see Unable to connect, this is not a problem
- Click Edit
- Search and select the matching product dataset
- Select "create" and allow the report to refresh.
 |   |
Once this is complete you should end up with a model with data connected live and the options illuminated to begin adding other sources.
Step 4: Add Additional Data Source
Finally you're ready to add your own data into this model.
- Click get data at the top
- Select your datasource.
- Once you select your data source, click "add a local model" to agree.
- Dataset will be converted to an AS model. You should see this pop up.
- Authenticate according to your selected source. Most commonly users want to join an Azure Sql Attribute, Excel File, or another Power BI Dataset. See our addendum at the bottom for specific examples. We'll skip by the specific options depending on your source selection.
- After authenticating click okay to this warning.
Step 5: Modeling and Publication
Once you've imported all sources, you can build new modeling on top of the combined sources. At the very least this usually involves relating the new tables. Most commonly, bringing in User data will require a join on Email or UPN as shown below. Once finished publish your work and verify your additional sources are refreshing. The Power BI datasets will not need a refresh as they are connecting live back to the dataset.
- Select relationship mode
- Create relationship.
- Add new filters or pivot on new attributes.
- Publish to the Power BI Service when you're finished. You may need to configure the refresh for items that are outside of the service.
Summary
If you're looking to combine data into tyGraph, Comp models in Power BI are the best way to achieve this.
At tyGraph we highly recommend this best practice because it maintains the data in Power BI Service making for a more secure deployment, that is centrally controlled, and much easier to maintain.