Which plot visualizes a continuous versus categorical variable and why?

WHAT THIS TESTS: This question checks whether you know the three seaborn categorical plot families and can select the right level of granularity for a continuous-versus-categorical comparison. Interviewers want to see that you think about the analytical goal before picking a chart type, rather than defaulting to a generic plot.

A GOOD ANSWER COVERS: First, name the appropriate plot types from the distribution or scatter families. For most exploratory work, boxplots or violinplots are ideal because they reveal median, quartiles, and outliers per category. If the dataset is small, a swarmplot can show every observation without overlap. Second, explain the specific insights you would hunt for: differences in central tendency between neighborhoods, variance or skew in house prices, the presence of outliers or bimodal distributions within a single category, and whether price ranges overlap or are cleanly separated. Third, mention ordering strategy, such as sorting neighborhoods by median price to make patterns immediately visible.

COMMON WRONG ANSWERS: A major red flag is suggesting a bar plot or line chart. Bar plots collapse the continuous distribution into a single estimate like the mean, hiding variance, multimodality, and extreme values. Another red flag is recommending a simple scatterplot without jitter or a beeswarm algorithm, which causes overplotting because all points in one category share the same x-axis position. Saying you would use a histogram alone also misses the categorical comparison aspect.

LIKELY FOLLOW-UPS: The interviewer might ask how you would handle fifty neighborhoods instead of five, which could motivate switching to a boxenplot or faceted violinplots to avoid overcrowding. They might ask how you would overlay a hue dimension such as property type, which changes color grouping within each categorical bin. Another follow-up is how you would communicate the plot to a non-technical stakeholder, which tests whether you can trade statistical precision for interpretability.

ONE CONCRETE EXAMPLE: Imagine you plot neighborhood on the x-axis and house_price on the y-axis using a violinplot. You notice that Neighborhood A has a tight distribution around three hundred thousand dollars, while Neighborhood B shows a bimodal shape with clusters at four hundred thousand and eight hundred thousand dollars. This immediately tells you that Neighborhood B contains two distinct housing submarkets, a insight that a simple mean bar would have completely obscured.

Source: seaborn.pydata.org