QGIS Graphical Modeller

The QGIS Graphical Modeller packages sequences of processing steps into reusable models. Instead of running the same operations manually each time, you build a model once and run it on new data with a single click.

Models combine everything covered so far: joins, aggregations, spatial operations, and attribute manipulation. You can share models with colleagues so they can replicate your workflows without rebuilding them from scratch.

Tip

Keep in mind that this is an advanced tool and it will take some time to fully master its potential. For now, the main thing is to understand the concept of why it exists and how it can help on your project work.

Graphical Modeller Interface

To open the Model Designer, go to Processing → Model Designer.

You can also access this from the top of the Processing Toolbox panel.

Modeller Interface

Main parts

  • Algorithms: A condensed version of what you have in your Processing Toolbox
  • Inputs: The list of types of inputs the user needs to specify to run the model.
  • Model canvas: The area where you build your model by dragging and connecting tools.
  • Variables: (optional) A place to set local variables that can be accessed by any of the algorithms
  • Save and Export buttons: Save your model file (.model3) locally, export as image, or export as a Python script.
  • Edit Model Help: Document what your model does and how to use it.
Note

Exported Python scripts include QGIS-specific libraries. Python workflows are covered in later lessons.

When to Use the Graphical Modeller

  • Automating repetitive tasks: By creating models, you can automate repetitive GIS tasks, saving time and reducing the potential for human error.
  • Ensuring consistency: Models ensure that the same steps are followed every time, leading to more consistent and reproducible results.
  • Facilitating complex workflows: The modeller allows you to chain multiple processes together, enabling the execution of complex workflows with a single click.
  • Calculating and comparing metrics for scenario testing: Use the modeller to efficiently calculate and compare different metrics for various scenarios.

Best practice

The modeller is a powerful tool for sharing processes with colleagues. However, to do so successfully you should have the following concepts in mind.

Note

Some of these concepts apply not just for the QGIS Modeller but also more generally when developing and sharing code workflows.

Documentation

Think about the end user, which may be your future self. Describe the goal of the algorithm, the input parameters, the return output, and any considerations for reproducibility.

The Edit Model Help feature provides a structured format for documenting your workflow.

You can also add comments to each algorithm, similarly to commenting code in Python.

Example of a comment

Documentation makes it easier to identify and fix issues in your models and update them as needed.

Simplicity

Prefer several simple, flexible models over one complex model. Keep models generalisable so they can work with different datasets and formats.

Models that balance simplicity with usefulness are easier to maintain and stay relevant longer.

Sharing

Within your team

GitHub Discussions (or similar platforms) allow threaded conversations for asking questions, sharing ideas, and discussing improvements. This collaborative approach helps identify issues early and refine models based on feedback.

Externally

External sharing requires addressing licensing, documentation, and user feedback. If your team’s deliverables include QGIS models for external use, ensure these considerations are handled first.

Workflows in data

Many software platforms offer no-code workflow tools: ArcGIS ModelBuilder, CARTO Workflows, Tableau Prep, and Rhino + Grasshopper. These tools allow users to create complex data workflows through graphical interfaces without extensive programming knowledge.

Exercise

The workflows below come from previous lessons and can be transformed into reusable models.

  • Approximating Blocks:
    1. Dissolving Neighbourhoods
    2. Selecting Streets by Intersection
    3. Buffering Streets
    4. Dissolving Street Buffers
    5. Differencing Street Buffers from Boundary
    6. Multipart to Singleparts

Block Approximation model (Daniel L. Xu)
  • Cafeterias by Street:
    1. Buffering Streets
    2. Count Points in Polygon
    3. Join the Data

Build this model using the madrid_nbhds.gpkg and madrid_premises.gpkg datasets from the datasets page.

Process:

  • Open the Graphical Modeller (Processing → Model Designer)

  • Add Inputs:

    • Select Vector Layer:
      • Description : Streets
      • Geometry Type: Line
      • Leave ‘Mandatory’ as True
    • Again select Vector Layer:
      • Description : Premises
      • Geometry Type: Point
      • Leave ‘Mandatory’ as True

  • First Algorithm:

    • Select Buffer
      • Description: Buffer Streets
      • Input Layer: Select Model Input → Streets
      • Distance: 15 (metres, assuming EPSG:25830)
      • Click OK

  • Second Algorithm:

    • Select Join Attributes by Location (Summary)
      • Description: Count Premises in Buffer
      • Input Layer: Select Algorithm output → Buffer Streets
      • Join Layer: Select Model Input → Premises
      • Geometric predicate: intersects
      • Fields to summarise: select any field (e.g., fid)
      • Summaries to calculate: count
      • Click OK

  • Third Algorithm:

    • Select Join Attributes by Field Value
      • Description: Join Counts to Streets
      • Input Layer: Select Model Input → Streets
      • Table field: fid
      • Input Layer 2: Select Algorithm output → Count Premises in Buffer
      • Table field 2: fid
      • For the Joined layer output, click the dropdown and select Model Output to mark this as the final result
      • Name the output: Streets with Premise Counts
      • Click OK

  • Save your model locally. You now have a reusable workflow that buffers streets and counts nearby premises.

Challenges

NoteChallenge 1: City blocks model

Using the provided city blocks model, create approximate blocks for your city.

Pay attention to CRS settings. After running the model, discuss what could be improved.

NoteChallenge 2: Population aggregation model

Create a model that automates the steps in the Spatial Aggregations section.