Evaluates chart selection for exploratory data analysis across categorical and numerical variables. Strong answers explain pie charts, distributions, comparisons, relationships, missingness, and when visuals mislead.
##### Scenario
Choosing appropriate visualizations during exploratory data analysis
##### Question
What is a pie chart? When should you use and not use a pie chart? What types of graphs are most useful during exploratory data analysis and why? How would you use a graph to explore the relationship between two variables?
##### Hints
Think about categorical vs numerical data, composition vs comparison charts, and scatterplots or heat-maps for relationships.
Quick Answer: Evaluates chart selection for exploratory data analysis across categorical and numerical variables. Strong answers explain pie charts, distributions, comparisons, relationships, missingness, and when visuals mislead.
Exploratory Data Visualization: Choosing the Right Charts
You are performing exploratory data analysis on a dataset with a mix of categorical and numerical variables. Your goal is to understand distributions, composition, and relationships to guide analysis and modeling decisions.
Constraints & Assumptions
Choose charts based on the data type and analytical question.
Distinguish composition, comparison, distribution, trend, and relationship tasks.
Explain when a chart can mislead.
Include practical EDA workflow advice.
Clarifying Questions to Ask Guidance
What are the key variables and their types?
Are we comparing categories, tracking time, measuring distribution, or examining relationships?
Are there many categories, rare categories, or missing values?
Is the audience exploratory analysts or business stakeholders?
Part 1 - Pie Charts
What is a pie chart, and when should you use or avoid it?
What This Part Should Cover Guidance
Define a pie chart as a part-to-whole chart whose slices sum to a total.
Use it only for a small number of mutually exclusive categories with clear differences.
Avoid it for many categories, precise comparisons, time trends, negative values, or categories that do not sum to a meaningful whole.
Recommend bar charts when comparison accuracy matters.
Part 2 - Useful Graphs for EDA
What graph types are most useful during exploratory data analysis and why?
What This Part Should Cover Guidance
Include histograms, density plots, box plots, violin plots, bar charts, line charts, scatterplots, heatmaps, pair plots, and missingness plots.
Match each chart to distribution, comparison, trend, composition, relationship, or data-quality questions.
Discuss log scales, faceting, and aggregation choices.
Part 3 - Exploring Relationships Between Variables
How would you use a graph to explore the relationship between two variables?
What This Part Should Cover Guidance
Use scatterplots for numeric-numeric relationships with trend lines and outlier checks.
Use box or violin plots for numeric outcomes across categories.
Use grouped bars, heatmaps, or mosaic-style charts for categorical relationships.
Watch for confounding, overplotting, nonlinearity, and Simpson's paradox.
Follow-up Questions Guidance
What chart would you use for a highly skewed numerical variable?
