Skip to navigation
Kumo PredictData & Graphs

What types of data can Kumo ingest?

View as Markdown

Kumo AutoML supports data from AWS S3, AWS Glue Catalog, Snowflake, Databricks, Google Cloud BigQuery, and Azure Data Lake Storage. Local upload availability depends on the deployment and release.

S3, ADLS, and local uploads support Parquet and CSV data. Warehouse connectors read tables or views according to the connector’s capabilities.

Kumo preprocesses numerical, categorical, multi-categorical, ID, text, timestamp, and embedding columns. Column types are inferred and can be corrected during table configuration.

See Data Connectors for current source support and Column Preprocessing for supported column types.