To load a JSON file directly into a DataFrame (assuming records are at the top level):

To compute the mean AND maximum of a “revenue” column within each group:

A left merge of customers (1,000 rows) with orders (3,000 rows) returns 4,500 rows. The most likely explanation:

You merge customers (1,000 rows) with orders (5,000 rows) using how=”left”. The result has 6,200 rows. What is most likely?

Which function counts unique values per group?

To merge two DataFrames where the customer ID column has different names (cust_id in one, customer in the other), use:

In str.contains applied to a column with NaN values, why pass na=False?

A categorical column (“region”) has 5% missing values and the rest are roughly evenly distributed across four categories. Which strategy is most defensible?

The standard recommended Plotly continuous colormap (perceptually uniform, colorblind-friendly):

Trailing/leading whitespace in a string column is most cleanly removed with: