Landry Bodjona
Back to the datasheet

LB-05 Data Analyst

Answers business questions in plain language by writing, checking and running SQL on the company’s sales data.

LB-05

Data Analyst Evaluation board

Live

Ask Basalt & Bean's sales data a question in plain words. The analyst writes the SQL, six layers of checks decide whether it may run, and you get the SQL, the table, a chart and an explanation. Then try to make it delete data.

Ask a question

Choose a question
One number. The result is a single value, so there is no chart. What was our revenue last quarter?
A table of seven products and a bar chart. What was the revenue of each product last quarter?
Two numbers per row, which make a point chart. How many active subscriptions are there for each delivery frequency?
Twelve months in date order, drawn as a line. Show revenue by month for last year.
A ranking where the order of the rows matters. Which 5 products earned the most revenue last year?
A ratio between 0 and 1, not a percentage. What was the cancellation rate last quarter?
The hardest question: it compares what customers bought in three periods. Which coffees lost the most repeat buyers last quarter?
The question that is sent
What was our revenue last quarter?

One number. The result is a single value, so there is no chart.

This copy of the site cannot ask questions live right now.

Your text is sent to a language model through the AI gateway to write the SQL. Do not type personal data. The data is demo data, and the trace of a run holds counts and names only, never your words.

What the analyst knows

This is everything the model is told about the data, and so everything a query may use. The allowlist is built from the same file: a table, column, join or function that is not listed here is refused.

Reading the semantic layer…

The Scope

Start a run and its trace appears here, step by step.