Workflow Interface

Workflows in GOAT provide a powerful visual automation system for creating sophisticated spatial analysis pipelines. Instead of running individual tools sequentially, you can connect multiple analysis steps together using a drag-and-drop canvas, creating an automated data processing that eliminates repetitive manual work.

The workflows can be reused across different datasets and scenarios. Each workflow consists of different types of nodes connected by edges, enabling you to:

• Automate complex analytical pipelines: Chain multiple tools together where the output of one analysis automatically feeds into the next
• Create multi-source data workflows: Build sophisticated analysis processes that integrate multiple datasets and processing steps
• Document automated processes: Add text annotations to explain methodology and analytical decisions
• Execute with flexible automation: Run individual nodes, execute workflow segments, or automate entire pipelines
• Build reusable automation templates: Store workflows within projects for repeatability across different scenarios
• Leverage advanced automation features: Use workflow variables and custom SQL for sophisticated parameterized analysis

The visual canvas interface makes complex spatial analysis automation accessible to users at all technical levels while maintaining full documentation of the analytical process for reproducibility and collaboration.

1. Interface Components

The workflow interface consists of two main panels and the workflow canvas, providing an intuitive workspace for visual workflow construction.

Workflow Management and Project Layers panel

This panel is located on the left and it is divided into two sections:

Workflow Management

• New: Click + New and choose From scratch or From template to create new analytical pipelines
• Workflow List: Manage existing workflows with options to rename, duplicate, and delete

Project Layers

• Layer Tree: Read-only display of project's data layers. You can drag and drop them onto the canvas to build the workflow.
• Add a Layer: Add new layers to the project to use them in the workflow and map mode.

Workflow Canvas

Canvas Workspace

The canvas workspace is where you can drag and drop nodes, zoom, pan, and select elements. It contains several control areas:

Canvas View Bar: Located in the bottom left corner of the canvas:

• Zoom In: Increase the canvas magnification for detailed work
• Zoom Out: Decrease the canvas magnification to see more of the workflow
• Fit View: Adjust the view to display the entire workflow at once
• Lock: Prevent accidental changes: while the canvas is locked, nodes and text cards can't be moved, connected or deleted. Panning and zooming still work. Click Unlock to edit again

Toolbar Controls: Located at the middle bottom of the canvas:

• Select: Default cursor for selecting and moving nodes
• Text Card: Add text annotations to document workflow steps
• Redo: Restore the last undone action
• Undo: Reverse the last action
• Variables: Create and manage workflow variables for reusable parameters
• Run: Execute the entire workflow

Minimap: Located in the bottom right corner of the canvas, providing an overview navigator for complex workflows.

Data View Controls: Located at the bottom of the canvas. Select a node to activate the panel. It shows the data of that node's layer:

• Table: Opens the attribute table of the selected node's layer
• Map: Opens a map preview of the selected node's layer (only available if the layer has geometry)

Keyboard shortcuts

On macOS, use ⌘ instead of Ctrl.

Shortcut	Action
Delete or Backspace	Delete the selected node, text card or connection
Ctrl+Z	Undo the last change
Ctrl+Shift+Z or Ctrl+Y	Redo the last undone change
Ctrl + click	Add a node or text card to the selection
Arrow keys	Move the selected node or text card
Esc	Clear the selection

With the mouse, drag on an empty part of the canvas to pan, and scroll to zoom.

• Undo and redo don't apply while you type in a field, such as a text card or a tool setting.
• Moving a node or text card, by dragging it or with the arrow keys, is a change you can undo like any other.
• A version loaded over the canvas, such as a change by the AI assistant or the saved version after Reload, is one step too: Ctrl+Z takes it back (see Change a workflow with the AI assistant). After Reload, Ctrl+Z brings back your version, and it is then saved over the other change without asking again.
• While the canvas is locked with Lock, Delete doesn't remove anything.

Tools and Configuration panel

The right panel changes depending if there is a node selected or not. If no node is selected then the Tools and History panel will be visible. If a tool node is selected, the Configuration panel will appear, and if a dataset node is selected the Dataset panel will appear.

Tools and History panel

Tools Tab

This tab contains categorized tools available for workflow construction, similar to the Map Mode Toolbox. Drag and drop tools onto the canvas to add them to your workflow. The tools are organized into the following categories:

• Data I/O
  • + Add dataset: Create dataset nodes
  • Fetch data: Fetch OpenStreetMap or Overture Maps features from GOAT's copy of them, fresh on every run. The node has no input. Described below under Fetch data Node.
  • Save as dataset: Save workflow results as permanent datasets. Configure the Dataset name, toggle Add to project to automatically add the result to the project layer list, and enable Overwrite on re-run to replace the previously exported dataset each time the workflow runs instead of creating a new one. Connected directly to a dataset node, it saves the features that match the node's filter, in the dataset's style.

Good practice

Give each Save as dataset node a descriptive name and enable Overwrite on re-run when running the same workflow repeatedly. This keeps your project clean by avoiding duplicate layers after each run.

Fetch data Node

A Fetch data node needs no input layer. It fetches OpenStreetMap or Overture Maps features and fetches them again on every run, so the result follows the source.

1. Drag Fetch data onto the canvas and click the node. The panel shows a Source section with a Choose data button.
2. Click Choose data. This opens the same dialog as OpenStreetMap and Overture Maps in Add layer: pick the layer type, the area, and tag or attribute filters or SQL. See Layers for the details.
3. Click Use in workflow to save the choice in the node. The node takes a suggested name such as "OSM points". Edit reopens the choice.

Nothing is fetched yet, so the nodes after it can be set up before the first run. While nothing is chosen, the node shows an attention icon. At most 1,000,000 features can be fetched, and very large areas are refused.

• Accessibility Indicators: All tools available in the Accessibility Indicators section of the Toolbox
• Geoanalysis: All tools available in the Geoanalysis section of the Toolbox
• Geoprocessing: All tools available in the Geoprocessing section of the Toolbox
• Data Management
  • Join, Merge, and other data manipulation tools
  • Custom SQL: Advanced data processing with SQL queries
• Control: Add a Conditional node that routes the layer to a True or False path based on defined conditions. See Conditional.

History Tab

Here you can see:

• Execution Log: Previous workflow runs with timestamps and status
• Execution Details: Duration, success/failure status, and error messages
• Result Access: Links to previous workflow outputs

Configuration Panel (Tool Node Selected)

When a tool node is selected, the right panel displays the Tool Configuration panel. Configure all tool-specific parameters for the selected analysis. You can also use workflow variables within parameter fields for dynamic values.

Dataset Panel (Dataset Node Selected)

When a dataset node is selected, the dataset panel appears with two available tabs:

Source Tab: View metadata from the data source and access table and map views. You can also change the dataset assigned to the node from this tab.

An empty dataset node shows Add dataset, with two sources: From Project and My datasets. Once a dataset is set, use Change Dataset to pick another.

Filter Tab: Apply filters specific to the workflow without affecting the original layer.

2. Example use cases

• Accessibility Analysis Pipeline: Create catchment areas, intersect with population data, calculate accessibility indicators, and export results
• Site Suitability Assessment: Buffer constraints, perform spatial overlays, apply weighting factors, and rank suitable locations
• Multi-Source Data Integration: Join multiple datasets, apply spatial filters, aggregate statistics, and create comprehensive analytical outputs
• Quality Assessment Workflow: Validate data quality, check spatial relationships, generate validation reports using Custom SQL
• Comparative Scenario Analysis: Use workflow variables to run identical analyses with different parameters or datasets

3. How to use the workflow interface

Getting Started

Begin with simple 2-3 node workflows to understand the interface, then gradually build more complex analytical pipelines as you become comfortable with the system.

Creating Your First Workflow

1

Navigate to Workflows: Click on the Workflows tab in GOAT's main navigation to access the workflow interface.

2

Create a new workflow: Click + New in the left panel and choose From scratch to start with an empty canvas, or From template to start from a ready-made workflow (see Start from a template). A new workflow is added to the list with a default name. In a project without workflows, the canvas offers the same in its centre: New workflow, Start from a template when workflow templates are available to you, and a What's a workflow? link to this page.

3

Rename it: Open the workflow's more options menu and choose Rename to give it a name that reflects your analytical objective (e.g. "Urban Accessibility Analysis").

Building Your Workflow

1

Add Data Sources: Add data to your workflow by either dragging + Add dataset from the right panel's Tools tab onto the canvas, or by dragging layers directly from the Project Layers panel on the left. Configure the dataset node to reference your input data layers.

2

Add Analysis Tools: Browse the categorized tool sections and drag the required analysis tools to your canvas. Tools are organized by function (Accessibility, Geoprocessing, etc.).

3

Connect Workflow Elements: Create edges by dragging from output handles to input handles. GOAT automatically validates geometry compatibility.

4

Configure Parameters: Click on each tool node to set analysis parameters. Use workflow variables for flexible, reusable configurations.

5

Document Your Process: Add text annotation to explain your methodology and analytical decisions.

Executing and Managing Workflows

1

Execute Workflow: Click Run at the bottom of the canvas to execute the entire workflow from start to finish. It is greyed out while the workflow has errors (see Checking and running a workflow).

2

Monitor Execution: Watch the progress indicators and check the job status in the main interface for execution updates.

3

Review Results: Use the Table and Map buttons at the bottom of the canvas to inspect the selected node's results once the workflow completes.

4

Save Results: Add and configure export nodes to save important results as permanent datasets in your project.

Auto-Save Feature

Workflows automatically save changes as you build them. The system preserves all configurations, connections, and execution states. If a save fails, a Not saved notice appears at the top of the canvas; click Retry to save again.

If the workflow was saved somewhere else since you opened it, for example by someone else or in another tab, your changes don't replace that version. A notice at the top of the canvas says This workflow was changed elsewhere, with who changed it and when, and nothing is saved until you choose: click Reload to load the saved version, or Keep mine to save your version over it. If the AI assistant changed the workflow while you were editing it, the notice reads The assistant changed this workflow while you were editing. Panning and zooming never cause this notice.

Checking and running a workflow

GOAT checks the workflow while you build it and shows what still needs attention. Nothing it finds stops you from building or saving.

• Next to Run, a chip such as 2 issues counts everything that was found. Click it to open the list, grouped by node; issues that concern the whole workflow come first.
• A node with an issue shows a badge at its bottom right corner. Point at the badge, or tap it, to read its messages.
• An input that must be connected but is not gets an amber ring around its handle.
• In a node's settings, a message appears under the setting it concerns, and the node's other issues are listed at the top. Required settings are marked with an asterisk (*).

Click an issue in the list or on a badge to go to it: the canvas moves to the node, selects it and opens its settings at the setting concerned.

Level	Colour	What it means
Error	amber	The workflow cannot run yet, for example because an input is not connected, a required setting is empty or a field does not exist in the input dataset.
Warning	grey	Worth a look, but it never stops a run, for example a node that is not connected to any other node.
Run error	red	The node failed in the last run, and its badge shows the node's own error message. It goes away when you change the node or run the workflow again.

While the workflow has errors, Run is greyed out; pointing at it tells you how many are left, for example Fix 2 issues to run. Warnings and run errors never disable Run.

If the workflow could not be checked, for example because the connection dropped, a chip Not checked appears. Click it to check again; a run is always checked when it starts.

Damaged workflows

If the saved version of a workflow cannot be read, for example because it was changed outside GOAT, it opens with the notice This workflow cannot be opened in place of the canvas, listing what is broken. Such a workflow cannot be edited or run there, and GOAT does not save over it. The other workflows of the project keep working.

Start from a template

1

In the Workflows panel, click From template to open the template browser, filtered to workflow templates.

2

Search or browse, select a template to preview it, and click Use template.

3

The workflow is added to your project. Where the template left an input open, pick a layer once it is in the project.

Build a workflow with the AI assistant

You can also describe an analysis to the AI assistant and have it build the workflow. In a project you can edit, open the assistant with Ask AI in the upper bar, in the Map or Workflows view, and describe what you need, for example Build a workflow that finds the areas within a 15-minute walk of the schools and counts the residents in them.

• When a workflow template fits, the assistant uses it. Otherwise it builds the workflow from the tools of your Toolbox, checks it with the same check as the editor (see Checking and running a workflow) and fixes what it can. It creates the workflow only when what is left is yours to choose: the layers of the steps it left open, and the settings that read them, such as a field.
• When you have to choose, for example which dataset to use, the assistant asks you in a card with one or more questions, at the bottom of the assistant in place of the message field. Several questions appear one at a time: choose an option (or several, where the question allows it), or pick Type your own answer and enter it; then click Next. Review then lists your answers, and Send answer sends them. A single question has Send answer right below it. The option the assistant suggests is marked Recommended and comes first. To answer in the message field instead, click Reply in text instead.
• The assistant adds a dataset from the catalog to your project only when you choose it as an option in such a card, not when you name it in an answer you type. A template is the exception: the datasets it ships come into your project with it (see Datasets in a template).
• If you choose to leave a dataset open, or no fitting data exists, the assistant leaves the dataset node empty and names the step for what is missing. The node then shows No dataset and Pick a layer, and the issue list shows No dataset is chosen for this step. under the step's name. Run stays greyed out until you have picked a layer and filled in what the issue list then asks for in the steps that use it.
• If you are still in the project and the workflow you have open has no unsaved changes, the new workflow opens in the editor at once; otherwise click Open on the workflow card in the answer.

Change a workflow with the AI assistant

You can also ask the assistant to change a workflow of the open project, for example Set the walking time to 10 minutes and add a note that explains it.

• The assistant reads the workflow first and changes only what you asked for: steps, settings, connections and text cards. It never deletes a workflow.
• It checks the changed workflow with the same check as the editor and doesn't save a change that adds an error. Its answer says what it changed.
• If the workflow is open in the editor, the canvas shows the change at once. Ctrl+Z (⌘+Z on macOS) in the canvas takes the change back as one step.
• An answer that changed a workflow shows Undo below it; on older answers, point at the answer to see it. Click it to take back all the changes of that answer, also when the workflow isn't open. The button then reads Undone.
• If the workflow was saved again after the answer, Undo asks first and says who changed it and when. Undo anyway takes back those later changes too. If the editor has changes that aren't saved yet, Undo asks as well; after Undo anyway, the notice at the top of the canvas lets you choose which version to keep.

Run a workflow with the AI assistant

• When you ask for a result, for example How many residents live within a 15-minute walk of a school?, the assistant builds the workflow and runs it. A workflow it created in this conversation runs without asking. Any other workflow, such as one you built yourself, runs only after you agree: the assistant asks in a card whether to run it, and nothing runs until you choose to.
• While the workflow runs, the answer shows a line such as Workflow running: step 2 of 4, Catchment Area with the time since it started. Open in Workflows opens the workflow, where the canvas shows the run. When the answer shows the workflow's card, Open on the card does this instead.
• The assistant waits for the run and reports the result in its answer, with the layers the run saved. If the run failed (Workflow run failed), it names the step that failed and what to change.
• If the run takes longer than the answer can wait, the line reads Workflow was running when this answer ended. The run goes on; ask the assistant again later for the result. If you stop the answer, a run it started stops too.

Save a workflow as a template

Once a workflow is set up, you can save it as a template so it can be reused, without repeating the setup, on other data or by other people.

1

Open the workflow's more options menu and choose Save as template….

2

Give the template a Name and, optionally, a Description and categories.

3

Choose the Location (a space and folder) where the template is saved. Everyone who can open that folder can use the template.

4

In the Inputs table, decide for each input layer whether it ships with the template or the template asks for a layer when used. A layer that ships is shared with the template so others can run it. For a shipped layer you can also choose Copy for every user (with content) or Empty copy for every user, so that everyone gets their own editable copy instead of your read-only layer (see Datasets in a template).

5

Click Save.

Your styling comes along

If you have run an export node and styled its result layer, the template keeps that style. People who use the template get your styling on their first run. An export that overwrites an earlier result keeps that layer's own style.

Update instead of duplicate

If a template was already saved from this workflow, you can update it from the source with the current version instead of creating a duplicate. Because the inputs are part of the saved snapshot, use Update template from source to change them.

4. Running workflows from the Map view

Workflows can also be run directly from the Map view without opening the workflow editor. Open the Toolbox, click the Workflows tab, and select a workflow from the list. If the workflow has variables, a Variables section appears where you can set values before running.