Background:

 

@RISK generates a SQLite database containing structured business data. 

While Power BI can load this database directly, Python visuals allow you to:

• Use plotting libraries such as Matplotlib.
• Produce custom chart types unavailable in native Power BI
• Apply advanced statistical or analytical functions.
• Highlight patterns with customized formatting and annotations.
• Build visuals that adapt dynamically to report filters and slicers.

A Python visual:

• Receives a data frame created from the fields you drag into the Python visual object.
• Runs your Python code to generate a chart.
• Renders the output as a static image inside Power BI.

 

Using Python Scripts in PowerBI with @RISK results

1. Load your SQLite database into Power BI

Import the SQLite tables with simulation results.

1. Go to Home > Get Data > More.
2. Select Other > ODBC.
3. Browse to @RISK’s generated .db file.
4. Select the tables you need (fact tables, dimensions, metadata).
5. Load the data into the model.

Note: Power BI will treat the tables normally, and you can build relationships between them as needed.

We recommend using ODBC to get data from the SQLite database but you can also write a Python script to read the SQLite file in Power BI.

2. Enable Python visuals in Power BI

Ensure Power BI is configured to use your local Python installation.

1. Open Power BI Desktop.
2. Go to File > Options and settings > Options.
3. Select Python scripting.
4. Under Python home directory, pick your Python installation folder.
5. Click OK

If Python is not installed, go to python.org and download it.

• During installation, make sure to include pip and add Python to PATH.

3. Add a Python visual

To insert a Python visual:

1. Click the Python visual icon in the Visualizations pane.
2. Power BI will prompt you to enable script visuals — accept the prompt.
3. A blank Python visual appears with a script editor below it.

Power BI will pass data into the Python script as a pandas dataframe called dataset.

Important: Make sure to install the pandas library by running the command: pip install pandas

4. Add data fields to the Python visual

Drag the fields you want to use into the Values section of the visual.

For example:

• Simulations.SimName
• Outputs.OutputName
• OutputData.OutputDataValue

Power BI will automatically build a dataframe containing only these fields.

Important:
Python visuals only receive the data you drag into the visual — not the entire table.

5. Write your Python code

When a Python visual is selected, the Python script editor becomes visible so you can write your code.

• For more information, see the Python script included in the @RISK example file called Advanced Modeling 5 – Python Visuals Dashboard.pbix

Important: The example Power BI file uses NumPy and Matplotlib. Make sure to install those libraries (pip install numpy matplotlib).

 

Best Practices

1. Keep visuals lightweight

Python visuals render as images. For better performance:

• Avoid heavy loops or large computations.
• Pre-aggregate data in Power BI whenever possible.
• Use smooth, simple plots where feasible.

2. Only pass necessary columns

Python visuals receive the fields you drag into them. To avoid performance bottlenecks:

• Send only the columns needed for the visual.
• Avoid including large text fields.
• Prefer pre-filtered tables or measures.

3. Python environment consistency

• Use a stable Python installation.
• Avoid upgrading plotting libraries without testing.
• Ensure pandas, NumPy and Matplotlib are installed and working.

4. Remember limitations

Python visuals:

• Render as static PNG images.
• Do not support interactivity from Python libraries.
• Refresh only when the dataset refreshes.

5. Test scripts outside Power BI

Debug your code in a standalone IDE (VS Code, PyCharm, Jupyter) before pasting into Power BI for faster iteration.

 

 

Related Article

https://learn.microsoft.com/en-us/power-bi/connect-data/desktop-python-visuals