Skip to content

How to Prepare AI Alias

This guide explains how to prepare an alias file and use it as the AI-inferred alias for a factor layer in CartoScope.

Concepts

0. Explore the Data with AI Assistance

You can explore your spatial dataset using the available AI-assisted tools within the platform.

Leveraging the current AI agent, you might:

  • Identify meaningful feature groupings or patterns
  • Inspect how factors are distributed spatially
  • Get inspiration for intuitive and consistent alias naming

1. Prepare Your Alias File

Alias File Format

  • The file must be a tab-separated file (.tsv) with one row per factor.
  • It should contains two columns:
    • 1st column (header: either index or name): numeric factor indices.
    • 2nd column (header: alias): aliases names.

Example Alias File

1
2
3
4
index   alias
1   Tumor
2   Stroma
3   Immune

Prepare Alias File via CartoScope

prep-alias-via-cartoscope

CartoScope offers a solution to prepare your alias file during exploration.

  1. Click "Load Layers" button to load the factor layer you want to annotate.
  2. Open the factor drawer to display your target factor layer.
  3. Review the Overview table to determine whether existing alias sources (e.g., curated or AI-inferred aliases)) are available.
  4. To annotate a factor, double-click the corresponding cell and enter your desired label.
  5. Once all annotations are finalized, click the “Alias” button at the top of the Overview table and select “Export alias” to generate and download the .tsv file. prep-alias-via-cartoscope2

If you want this alias file to be available as the AI-inferred alias for a factor layer of a dataset, please:

  1. specify it in catalog.yaml,
  2. upload the file to the AWS S3 directory or Zenodo repository that hosts this dataset.

Example catalog.yaml with Alias Files

In the example, we want to provide the ai_inferred_alias.tsv file to the target_layer factor layer in the example_data dataset.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
id: example_data
title: example_data
...
assets:
...
factors:
    - id: target_layer
    name: target_layer
    ...
    alias_ai: ai_inferred_alias.tsv

In this setup:

  • alias_ai is the optional AI-generated alias table used as the AI-inferred source.