Connect your data

Bring data into Golden Analytics from files, Google Sheets, or a cloud warehouse such as Snowflake, BigQuery or Salesforce Reports.

3 min readUpdated October 28, 2026

Connect is the first stage of an analysis. You pick a source, and Golden loads it into a dataset you can explore, prepare, and visualize.

Where to start

From the sidebar, open Datasets to create a Shared Dataset, or start a new analysis to load data into a session.
On the Connect screen, pick a source category from the left nav: Cloud data sources, Upload files, Shared datasets, or Sample datasets.
Follow the prompts for that source. Once data loads, you move on to Prepare and Discover.

Upload files

The file picker accepts .csv, .txt, .tsv, .xls, .xlsx, .json, and .parquet. Multi-sheet workbooks prompt you to pick a sheet before loading.

Note: Some Excel files can't be read, and Golden refuses them with advice on screen rather than a generic error. A password-protected .xlsx needs the password removed in Excel before you upload a copy. An older Excel 97-2003 .xls file has to be opened in Excel and re-saved with File → Save As as .xlsx or .csv. And a workbook whose single sheet holds more data than can be opened in one piece asks you to save each sheet you need as its own CSV and upload those instead. The message tells you which case you hit.

Google Sheets

Sign in with Google, pick a spreadsheet, then pick the tab you want. Golden imports the chosen sheet as a dataset. Note that by default, Golden shows only the content that you own in Google Sheets. If you want to access content that was shared with you by another person, select the "include shared with me" toggle.

Cloud data sources

Under Add new connection, pick Snowflake, Databricks, Google BigQuery, Amazon Redshift, Microsoft Fabric, Azure SQL Database, or Salesforce Reports. These cloud connectors are available to every workspace.

Note: When you save a cloud connection, Golden verifies it by running the same check as the Test Connection button. This warns but never blocks: if the connection can't be reached, the reason shows inline as "Last connection test failed" and the Save button becomes "Save anyway", so a second click saves it regardless. Editing any field re-arms the check so the next save tests again. For Snowflake, the test also confirms your chosen warehouse is one your role can access.

Tip: Saved connections are scoped to a workspace, plus any Global (workspace-less) connections. Anyone in the space can re-use them without re-entering credentials.

Shared and sample datasets

The Sample datasets tab with built-in datasets to try the product: Golden Records, E-Commerce Sales, Financial Portfolio, Marketing Campaigns, and HR Workforce. — Sample datasets — try the product with built-in data before connecting your own.

Shared datasets are curated tables your teammates saved earlier — they appear under the Shared datasets tab with a count badge. Sample datasets are built-in CSVs you can use to try out the product before connecting your own data.