Category: Semantic Link Labs

Find Unused Objects In Power BI Semantic Models With Semantic Link Labs

A few weeks ago I wrote about how Semantic Link Labs now has tools with interactive UIs and showed how you could use this to view lineage and so Vertipaq-Analyzer-stuff in a notebook. I didn’t show what I think is the coolest new feature though: the ability to find the tables, columns and measures in a semantic model that aren’t used and which therefore could be deleted. There are tons of excellent third-party tools that do this available already of course but the advantage of using Semantic Link Labs for this is that you can automate the process of finding unused columns and run it from a notebook and, crucially, you have two ways of finding those unused columns: from analysing the structure of downstream reports and from analysing DAX queries captured in Workspace Monitoring.

To test this I used a simple semantic model with one table in a workspace with Workspace Monitoring enabled:

I then created a report with a single visual showing the Person Count measure broken down by gender:

Next I created a notebook with the following code that uses find_unused_objects:

%pip install semantic-link-labs
import sempy_labs as labs
labs.semantic_model.find_unused_objects(
    dataset = "insertsemanticmodelidhere",
    workspace = "insertworkspaceidhere",
    method = "WorkspaceMonitoring",
    visualize=True
)

This displayed the following widget which queried Workspace Monitoring and found three DAX queries in the previous seven hours:

Clicking the Analyze button showed me – as expected – that only the Gender and Person Count measures had been used in these three queries and that all the rest of the columns were unused:

Of course you don’t need to use the widget and can get a pandas dataframe back instead.

Tools With Interactive UIs In Fabric Notebooks With Semantic Link Labs

There’s so much going on in the Fabric community that it can be hard to keep up with it all. Semantic Link Labs is a great example: in the six months or so since I last had a proper look at it my colleague Michael Kovalsky has done a whole load of cool things and it wasn’t until I had a chat with him recently that I realised how much had changed. Most importantly, for someone old-fashioned like me who still likes tools with a UI, a lot of new functionality has been added which has a UI and is usable with minimal coding.

To illustrate this, I created a new Fabric notebook in a workspace and installed Semantic Link Labs:

%pip install semantic-link-labs

I then headed over to the Semantic Link Labs Code Examples and copied some of the code from there into cells in my notebook. For example, the following code:

import sempy_labs.semantic_model
sempy_labs.semantic_model.lineage_view()

…opened up a tool for exploring report and semantic model lineage appearing within the notebook. I could connect to a semantic model:

…and then see which reports are connected to it and even look for broken visuals in those reports:

There’s also a version of Vertipaq Analyzer:

import sempy_labs as labs
dataset = 'insert id or name of semantic model here'
workspace = 'insert id or name of workspace here'
x = labs.vertipaq_analyzer(dataset=dataset, workspace=workspace)

…and a whole load of other things which probably deserve their own blog post. So if, like me, you had assumed that Semantic Link Labs was for people who like writing code rather than using a UI, take another look – you’ll probably find something useful.