Upload Your Dataset

File Requirements

• Format: Parquet files only
• Structure: Single table with one row per observation
• Data quality: Remove invalid records (e.g., negative exposures) before upload

Upload Process

1. Navigate to the dataset upload area
2. Drag and drop your Parquet file or click to browse
3. To use an existing dataset, select from your previously uploaded files (stored in S3)

Your datasets are private—only you can access them.

After Upload

Verify the data preview shows:

• Your target column
• All intended predictor variables
• Correct data types for each column

Train/Test Split

Choose one of:

• Single file: System automatically splits into train/test
• Separate files: Upload training and test datasets separately

Next Steps

Configure your model