Data Literacy and Data Science Fundamentals
What the exam tests
This domain tests whether you can identify data types and scales, describe how data is collected and cleaned, choose an appropriate chart, and interpret descriptive statistics correctly rather than quoting a mean blindly.
Core concepts
- Structured versus unstructured: tables and spreadsheets are structured; text, images, and audio are unstructured.
- Measurement scales: nominal (labels, no order), ordinal (ordered labels), interval (equal gaps, no true zero), ratio (equal gaps with a true zero).
- Collection: surveys, sensors, web logs, transactions, and open datasets. Sampling must be random and representative to avoid bias.
- Cleaning: remove duplicates, handle missing values (drop, or impute with mean/median/mode), correct inconsistent formats, and treat outliers after checking whether they are errors or genuine.
- Visualisation: bar chart for category counts, histogram for a numeric distribution, scatter plot for the relationship between two numeric variables, box plot for spread and outliers, line chart for trends over time.
- Descriptive statistics: mean, median, mode, range, variance, standard deviation, quartiles, and correlation.
Worked example
Five salaries in lakhs: 20, 22, 25, 28, 100. Sum = 195, so mean = 39. Sorted, the middle value is 25, so median = 25. The 100 is an extreme outlier that pulls the mean far above four of the five values. Because the median resists outliers, reporting median = 25 describes the typical salary far better; reporting mean = 39 would mislead.
Traps that cost marks
- Confusing ordinal with nominal, or interval with ratio.
- Using a pie chart for many categories, or a line chart for unordered categories.
- Saying correlation proves causation.
- Saying standard deviation is a measure of centre; it measures spread.
- Imputing missing values with the mean when the column is skewed; the median is safer.