Configure Data Source and Data Preparation

Configuring a data source is selecting the relevant tables and columns for the AI to interpret. It is important to select all the necessary tables and columns, but keep it minimal to improve processing speed.

If you have not set up a project, click here.

Data Table Selection

• ‍Start by identifying and selecting the data table(s) you wish to analyze. You can select any table under the raw tables tab. Make sure you select only the relevant tables and columns that you will ask about for optimal chat experience.

Column Curation

• Within each table, carefully review and pick out columns that are directly related to the questions you intend to ask.

We recommend that you select up to 15 columns. You can choose more than 15 columns but we recommend keeping it minimal to avoid clutter and maintain focus on relevant data points.

Repeat and continue the process of selecting tables and refining columns iteratively. With each iteration, refine your data selection for relevance and precision, which is instrumental in building an efficient dataset for querying.

Advanced Data Manipulation:

• For users comfortable with SQL commands to join tables, clean data, and perform complex manipulations. You can apply this feature to handle large datasets when you require detailed analytical functions, suitable for those with expertise in data analysis. See Datasets for more information.