CompTIA Data+ flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

Ordinal data

Ordinal data is categorical data whose categories carry a meaningful order or rank, such as low, medium, and high, but the distances between ranks are not equal or measurable.

Ratio vs interval scale

Ratio data has a true zero so ratios are meaningful, as with revenue or weight. Interval data has equal spacing but no true zero, so 20 degrees Celsius is not twice 10 degrees.

Fact table

In dimensional modeling, the fact table stores quantitative measures plus foreign keys that link to dimension tables, which hold the descriptive attributes used for filtering, grouping, and labeling.

Star schema

A star schema places one central fact table around denormalized dimension tables, producing fewer joins and faster queries than a snowflake schema, which normalizes dimensions into multiple related tables.

Data lake

A data lake stores raw structured, semi-structured, and unstructured data at scale in its native format, applying schema on read rather than enforcing a schema when the data is written.

Third normal form

Third normal form requires a table to be in second normal form and to contain no transitive dependencies, meaning every non-key column depends only on the primary key, not on other non-key columns.

Semi-structured data

Semi-structured data, such as JSON or XML, uses tags or keys to organize values but does not fit a rigid tabular schema, so it is typically queried with schema-on-read tools.

CRISP-DM

CRISP-DM is a six-phase data mining methodology: business understanding, data understanding, data preparation, modeling, evaluation, and deployment. The process is iterative and often loops back to earlier phases.

Precision vs recall

Precision is true positives divided by all predicted positives; recall is true positives divided by all actual positives. Precision penalizes false positives, while recall penalizes false negatives.

F1 score

The F1 score is the harmonic mean of precision and recall, calculated as 2 times precision times recall divided by precision plus recall, and is preferred when class distribution is imbalanced.

k-fold cross-validation

k-fold cross-validation splits data into k equal folds, trains on k minus one folds and tests on the remaining fold, repeating until every fold has served once as the test set.

Overfitting

Overfitting occurs when a model learns noise specific to the training data, producing very low training error but poor performance on new data. Remedies include more data, simpler models, and regularization.

Association rule lift

Lift measures how much more often items A and B appear together than expected if they were independent. Lift above 1 indicates positive association, exactly 1 independence, and below 1 negative association.

Stratified sampling

Stratified sampling divides the population into subgroups such as classes or regions and samples proportionally from each, preserving the population composition and reducing bias when the dataset is imbalanced.

Median vs mean

The mean is the arithmetic average and is pulled toward outliers and skew; the median is the middle value of sorted data and is preferred for skewed distributions such as income.

Type I error

A Type I error is a false positive: rejecting a null hypothesis that is actually true. Its probability equals the significance level alpha, commonly set at 0.05 or five percent.

p-value

A p-value is the probability of observing a result at least as extreme as the sample data, assuming the null hypothesis is true. When it falls below alpha, the null is rejected.

Coefficient of determination

R-squared, the coefficient of determination, is the proportion of variance in the dependent variable that the regression model explains. It ranges from 0 to 1, with higher values indicating a better fit.

Descriptive vs diagnostic analysis

Descriptive analytics summarizes what happened using aggregates and trends, while diagnostic analytics explains why it happened using drill-down, correlation, and root-cause techniques such as the five whys.

Chi-square test

The chi-square test of independence determines whether two categorical variables are associated by comparing observed frequencies against the frequencies expected if the variables were independent of each other.

Inner join

An inner join returns only rows with matching keys in both tables, whereas a left join returns all rows from the left table plus matches, filling unmatched right-side columns with nulls.

Box plot components

A box plot displays the median, first quartile, and third quartile inside the box, whiskers extending to 1.5 times the interquartile range, and any points beyond the whiskers as outliers.

Truncated y-axis

Starting a bar chart's y-axis above zero truncates the scale and visually exaggerates differences between bars. Bar charts should start at zero to keep lengths proportional to values.

Diverging color scale

A diverging color scale uses two contrasting hues meeting at a neutral midpoint to show deviation above and below a reference value, such as variance from a target or budget.

Choropleth map

A choropleth map shades geographic regions by the value of a variable, using color intensity to encode magnitude. It requires normalized rates rather than raw counts to avoid distortion from population size.

Histogram

A histogram groups continuous numeric data into equal-width adjacent bins and plots the count or frequency per bin, revealing the distribution's shape, center, spread, and skew.

Data-ink ratio

The data-ink ratio is the proportion of a chart's ink devoted to actual data. Maximizing it means removing gridlines, borders, backgrounds, and other decorative elements that carry no information.

Data steward

A data steward is responsible for the day-to-day quality of a data domain, defining standards and resolving issues, while the data owner holds ultimate accountability for that same data.

Data quality dimensions

Common data quality dimensions are accuracy, completeness, consistency, timeliness, validity, and uniqueness. Each is measured against a defined threshold and monitored through recurring quality reporting.

Data lineage

Data lineage documents the origin, movement, and transformation of data from source system to final report, enabling impact analysis, root-cause investigation of errors, and audit compliance.

Role-based access control

Role-based access control grants permissions to roles rather than to individuals, and users inherit permissions by being assigned to a role, simplifying administration and enforcing least privilege.

Data masking

Data masking replaces sensitive values with realistic but fictitious substitutes so non-production environments can be used for testing without exposing personal data while preserving format and referential consistency.