Data Analysis and Visualization
The Data Analysis domain tests the data cycle: collecting, cleaning, organizing, analyzing, visualizing, and interpreting data. Candidates should choose appropriate summaries and charts and recognize misleading presentations.
Data types and tools
Data may be structured, such as spreadsheet rows and database tables, or unstructured, such as text and images. Spreadsheets support formulas, sorting, filtering, and charts. Databases use tables, records, fields, and queries. Cleaning addresses missing values, duplicates, inconsistent formats, and outliers-but removal must be justified, not automatic.
Statistics and charts
Mean, median, and mode describe center; range and standard deviation describe spread. The median resists outliers. Correlation does not establish causation. Bar charts compare categories, line charts show trends over time, histograms show distributions, and scatter plots show relationships. Pie charts work for parts of one whole and should total 100%.
Worked example
Scores are 70, 72, 75, 78, and 100. The mean is 79; the median is 75. The 100 is an outlier that pulls the mean upward. If the goal is a typical score, the median is more representative. To show the distribution of many scores, a histogram is better than a pie chart because pie slices imply parts of a whole.
Traps
Do not use a pie chart for overlapping or many-category data. Watch truncated axes and inconsistent scales. Do not claim causation from correlation. Do not delete outliers merely because they are inconvenient.