Practice warehouse
Ask the course's sales data questions, here in your browser. You ask in SQL, a short language for asking a table a question. One question in SQL is called a query.
Nothing to install and no account. Your queries run on your own computer. The tables have the same names as in the course's BigQuery dataset, so you can use the same SQL in the class notebook.
Your query
- Starting the practice warehousewaiting
- Downloading the course datawaiting
- Getting the tables readywaiting
The first time can be slow on a weak connection. You can type while you wait.
Examples to start from
Each one shows what a query for that session looks like. They are examples, not the answers to your hand card. Each one runs here and in BigQuery without a change.
No SQL yet? Build a prompt and let Gemini write it. Ran the AI's query? Match its numbers against your hand card.
Results
Run a query to see its results here.
Match your hand card
Type the numbers from your hand card, and the numbers the AI's query printed for the same rows. The checker says, for each number, whether yours, the AI's, or both disagree with the answer key. It never shows the key.
What you type stays in this browser only. Nothing is sent anywhere.
Chart
Draw your results. The course rule: a chart's title states its message, not just its topic. Check three of its points against your hand card.
Ask Gemini to write the SQL
Write your steps in the five boxes: the same five parts as your prompts in class. The page adds the table names and their columns for you. Then copy the prompt, paste it into Gemini, and paste Gemini's SQL back into the query box above.
The page adds: the two table names, every column with its type and meaning, and a request for one SQL query for BigQuery.
See the whole prompt
Paste nothing personal into any AI tool. Use course data only: no names, emails, phone numbers or customer lists from work.
When Gemini answers, copy only the SQL. It usually sits in a grey box that starts with SELECT or WITH. If its result and your hand card disagree, follow the mismatch protocol.
The data: Online Retail II by Daqing Chen, UCI Machine Learning Repository, licence CC BY 4.0, doi:10.24432/C5CG6D. The course adds the row column and the cleaned table course.sales. BigQuery scan sizes come from BigQuery's own estimates for each column of the course tables.