Getting Started with Potato in 5 Minutes

Ready to start annotating? This quick tutorial will have you labeling data in under 5 minutes. We'll create a simple sentiment analysis task to demonstrate the basics.

Step 1: Install Potato

First, install Potato using pip:

pip install potato-annotation

Verify the installation:

potato --version

Step 2: Create Your Configuration

Create a file called config.yaml with this minimal configuration:

annotation_task_name: "Quick Start Sentiment Analysis"
 
# Your data file
data_files:
  - data.json
 
# Annotation interface
annotation_schemes:
  - annotation_type: radio
    name: sentiment
    description: "What is the sentiment of this text?"
    labels:
      - Positive
      - Negative
      - Neutral

Step 3: Create Sample Data

Create a data.json file with some sample texts:

{"id": "1", "text": "I love this product! It's amazing!"}
{"id": "2", "text": "Terrible experience. Would not recommend."}
{"id": "3", "text": "The weather is partly cloudy today."}
{"id": "4", "text": "Best purchase I've ever made!"}
{"id": "5", "text": "This is the worst service I've encountered."}

Step 4: Start Annotating

Launch the annotation server:

potato start config.yaml

Open your browser to http://localhost:8000. You'll see the annotation interface!

Step 5: Label Your Data

1. Log in with any username (Potato creates accounts automatically)
2. Read the text displayed
3. Select the appropriate sentiment label
4. Click "Submit" to save and move to the next item

Step 6: Export Your Annotations

Your annotations are automatically saved to the annotation_output/ folder. Each annotator's work is saved in a separate file.

To view your annotations:

cat annotation_output/your_username.jsonl

What's Next?

Congratulations! You've completed your first annotation task. Here's where to go from here:

• Add keyboard shortcuts: Speed up annotation with hotkeys
• Customize the interface: Add instructions, tooltips, and validation
• Set up multiple annotators: Configure user management and quality control
• Explore annotation types: Try spans, checkboxes, Likert scales, and more

Check out our documentation for detailed guides on each feature, or browse the showcase for complete configuration examples.

Having trouble? Check our FAQ or open an issue on GitHub.