Understanding Data Representation and Plotting in Biostatistics Course

Lesson 1 • Sources and Quality of Biomedical Data

Surveys clinical trials, registries, surveys, and omics datasets as data sources. Evaluates reliability, completeness, and bias risk before analysis begins.

Lesson 2 • Missing Data Identification

Defines MCAR, MAR, and MNAR mechanisms and their impact on representativeness. Introduces visual and tabular tools for detecting and documenting missingness patterns.

Lesson 3 • Dataset Structure and Organization

Explains tidy data principles, rows-as-observations format, and wide vs. long layouts. Proper structure prevents errors in all downstream visualization and analysis steps.

Lesson 4 • Types of Variables in Biology

Distinguishes continuous, discrete, nominal, and ordinal variables with biomedical examples. Establishes the vocabulary needed for all subsequent data handling decisions.

Lesson 1 • Distribution Shape and Skewness

Quantifies skewness and kurtosis and links shape to appropriate statistical methods. Recognizing non-normality guides transformation and non-parametric test selection.

Lesson 2 • Summarizing Data for Publication

Formats Table 1 style descriptive summaries following reporting standards. Applies correct notation for means, medians, and proportions with appropriate precision.

Lesson 3 • Frequency Tables and Cross-Tabulation

Constructs absolute, relative, and cumulative frequency tables for categorical data. Cross-tabulation reveals associations between two categorical variables before formal testing.

Lesson 4 • Measures of Spread and Variability

Covers variance, standard deviation, IQR, and range with biomedical interpretations. Variability metrics directly inform confidence interval width and sample size planning.

Lesson 1 • Choosing the Right Chart Type

Maps variable types and analytical goals to appropriate chart families. Correct chart selection is the single most impactful visualization decision.

Lesson 2 • Color, Contrast, and Accessibility

Applies colorblind-safe palettes, contrast ratios, and redundant encoding for accessibility. Accessible design ensures findings reach all readers without misinterpretation.

Lesson 3 • Axes, Scales, and Labels

Covers axis origin, scale breaks, tick spacing, and annotation best practices. Proper axis design prevents distortion and ensures readers interpret magnitude correctly.

Lesson 4 • Avoiding Misleading Visualizations

Identifies truncated axes, dual axes, cherry-picked ranges, and 3-D distortions as deceptive practices. Students audit existing biomedical figures for integrity violations.

Lesson 1 • Q-Q Plots for Normality Assessment

Constructs quantile-quantile plots and interprets deviations from the reference line. Q-Q plots guide the choice between parametric and non-parametric methods.

Lesson 2 • Histograms and Bin Selection

Builds histograms and applies Sturges, Scott, and Freedman-Diaconis rules for bin width. Bin choice critically shapes the apparent distribution of biological measurements.

Lesson 3 • Violin Plots and Hybrid Displays

Combines density estimation with box plot elements to show full distributional shape. Hybrid plots reveal multimodality hidden by box plots alone.

Lesson 4 • Kernel Density Estimation Plots

Explains bandwidth selection and kernel functions that smooth empirical distributions. Density plots complement histograms when comparing overlapping groups.

Lesson 1 • Visualizing Rates and Proportions

Plots incidence rates, prevalence proportions, and risk ratios with appropriate confidence intervals. Uncertainty display is mandatory for any proportion reported in biomedical research.

Lesson 2 • Mosaic and Spine Plots

Constructs mosaic plots to visualise two-way contingency tables with area-encoded proportions. Mosaic plots reveal interaction patterns between two categorical variables.

Lesson 3 • Pie and Donut Charts

Evaluates when pie charts are acceptable and when bar charts are superior for part-to-whole data. Limits on the number of slices and labelling strategies are emphasised.

Lesson 4 • Dot Plots and Cleveland Dot Charts

Applies Cleveland dot charts as a low-ink alternative to bar charts for ranked comparisons. Dot plots reduce visual clutter while preserving precise value reading.

Lesson 1 • Trend Lines and Smoothers

Adds linear regression lines, LOESS curves, and confidence bands to scatter plots. Smoothers reveal non-linear trends that straight regression lines obscure.

Lesson 2 • Pair Plots and Scatterplot Matrices

Builds scatterplot matrices to explore all pairwise relationships in a multivariate dataset. Diagonal panels display univariate distributions for each variable simultaneously.

Lesson 3 • Faceting and Small Multiples

Applies facet grids and facet wraps to display the same plot across subgroups simultaneously. Small multiples enable direct visual comparison while controlling for a third variable.

Lesson 4 • Scatter Plots and Overplotting Solutions

Constructs scatter plots and applies jitter, alpha transparency, and hexbin binning to handle overplotting. Scatter plots are the primary tool for exploring continuous variable relationships.

Lesson 1 • Survival Curves and Kaplan-Meier Plots

Constructs Kaplan-Meier curves with censoring marks, risk tables, and log-rank p-values. Survival plots are essential for time-to-event outcomes in clinical and epidemiological research.

Lesson 2 • Volcano Plots for Omics Data

Builds volcano plots combining fold-change on the x-axis with statistical significance on the y-axis. Volcano plots are the standard display for differential expression and GWAS results.

Lesson 3 • Confidence Intervals in Plots

Displays confidence intervals as error bars, line ranges, and shaded bands on various plot types. CI visualization replaces misleading standard error bars in modern biomedical reporting.

Lesson 4 • Effect Size Visualization

Plots Cohen's d, odds ratios, and relative risks with uncertainty to communicate practical significance. Effect size figures shift focus from p-values to clinically meaningful magnitudes.

Lesson 1 • Figure Export and Resolution Standards

Covers raster vs. vector formats, DPI requirements, and journal-specific figure specifications. Correct export settings prevent pixelation and colour shifts in published figures.

Lesson 2 • Version Control for Analysis Code

Applies version control concepts to track changes in plotting scripts and data processing steps. Version control enables collaboration and complete audit trails for regulatory submissions.

Lesson 3 • Dynamic Reports and Notebooks

Embeds code, output, and narrative in literate programming documents for integrated reporting. Dynamic reports eliminate copy-paste errors between analysis and manuscript.

Lesson 4 • Peer Review and Figure Audit

Establishes a checklist-based review process for evaluating figures before submission. Systematic auditing catches axis errors, missing labels, and accessibility failures.