Chapter 6 Field Specific Visualizations
6.1 Learning Objectives
6.1.1 Repeated Measures Plot (maybe?)
Description: Filler description
When to use:
- Type of data
- Number of variables
- Primary data encodings
- Used for ….
Strengths:
- Strength 1
- Strength 2
Weaknesses:
- Weakness 1
- Weakness 2
- Weakness 3
Alternatives:
- Alternative (if applicable)
- Alternative 2 (if applicable)
Example plot
Interpretation
6.1.2 Caterpillar Plot
Description: Displays numerical point estimates (such as odds ratios or coefficient effect estimates) as the caterpillar body and their variation or the confidence level as caterpillar legs. Caterpillar plots are practically the same as Forest plots except for one major difference: the categories are ordered by magnitude of the point estimates.
When to use:
- Numerical data & categorical data
- Several variables are required
- Primary data encodings are points and lines
- Used to find a ranking or for meta-analyses or comparing models or odds ratios
Strengths:
- Provides comparison to a reference line (e.g., baseline or null effect)
- Efficient readability since values are ordered by magnitude
- Displays variation both intra- and inter-study
Weaknesses:
- Cluttered and difficult to read if lots of studies or models are included
- A lot of context and information are in the axis labels
- Plot is just showing the effect estimates and uncertainty but no further context such as bias
Alternatives:
- Funnel plot
- Heatmap
Example plot
Interpretation
6.1.3 Waterfall Plot
Description: Bar charts which show treatment effect. The numerical axis is compared to a baseline. The bars are arranged or ordered according to magnitude and each bar typically corresponds to a different patient or sample.
When to use:
- Numerical data (may have categorical data for color)
- Minimum of 1 variable if the difference between patient/sample and baseline is already computed, but a minimum of 2 variables otherwise
- Primary data encodings are bars and position
- Typically used to show patient response to a treatment but the design can be used for other use cases around this theme
Strengths:
- Displays patient or sample response heterogeneity within one snapshot
Weaknesses:
- The snapshot compares two time points but not a range of time
- For tumor response, there’s an accepted notion of what above or below the baseline means but this might not be true for other applications
Alternatives:
- Spider plot incorporates time
- Swimmer plot incorporates time but loses the comparison to baseline
- Forest or Caterpillar plots can be use for treatment effect
- Kaplan-Meier curves can be used for outcome
- Heatmaps can be used to incorporate time
How could a heatmap be an alternative plot?
Each row would be a patient (y-axis) and the columns could be time. Use color to communicate the change from baseline. Perhaps consider faceting or arranging samples according to response category.
Example plot
Interpretation
6.1.4 Kaplan-Meier (KM) Curve (definitely)
Description: Filler description
When to use:
- Type of data
- Number of variables
- Primary data encodings
- Used for ….
Strengths:
- Strength 1
- Strength 2
Weaknesses:
- Weakness 1
- Weakness 2
- Weakness 3
Alternatives:
- Alternative (if applicable)
- Alternative 2 (if applicable)
Example plot
Interpretation
6.1.5 Volcano Plot (definitely)
Description: Filler description
When to use:
- Type of data
- Number of variables
- Primary data encodings
- Used for ….
Strengths:
- Strength 1
- Strength 2
Weaknesses:
- Weakness 1
- Weakness 2
- Weakness 3
Alternatives:
- Alternative (if applicable)
- Alternative 2 (if applicable)
Example plot
Interpretation
6.1.6 Enrichment Plot (maybe)
Description: Filler description
When to use:
- Type of data
- Number of variables
- Primary data encodings
- Used for ….
Strengths:
- Strength 1
- Strength 2
Weaknesses:
- Weakness 1
- Weakness 2
- Weakness 3
Alternatives:
- Alternative (if applicable)
- Alternative 2 (if applicable)
Example plot
Interpretation
6.1.7 MA Diagram (definitely)
Description: Filler description
When to use:
- Type of data
- Number of variables
- Primary data encodings
- Used for ….
Strengths:
- Strength 1
- Strength 2
Weaknesses:
- Weakness 1
- Weakness 2
- Weakness 3
Alternatives:
- Alternative (if applicable)
- Alternative 2 (if applicable)
Example plot
Interpretation
6.1.8 Principal Components Analysis (PCA) Plot (probably)
Description: Filler description
When to use:
- Type of data
- Number of variables
- Primary data encodings
- Used for ….
Strengths:
- Strength 1
- Strength 2
Weaknesses:
- Weakness 1
- Weakness 2
- Weakness 3
Alternatives:
- Alternative (if applicable)
- Alternative 2 (if applicable)
Example plot
Interpretation
6.1.9 Circos Plot (maybe)
Description: Filler description
When to use:
- Type of data
- Number of variables
- Primary data encodings
- Used for ….
Strengths:
- Strength 1
- Strength 2
Weaknesses:
- Weakness 1
- Weakness 2
- Weakness 3
Alternatives:
- Alternative (if applicable)
- Alternative 2 (if applicable)
Example plot
Interpretation
6.1.10 Biomedical Knowledge Graph (KG) (probably)
Description: Filler description
When to use:
- Type of data
- Number of variables
- Primary data encodings
- Used for ….
Strengths:
- Strength 1
- Strength 2
Weaknesses:
- Weakness 1
- Weakness 2
- Weakness 3
Alternatives:
- Alternative (if applicable)
- Alternative 2 (if applicable)
Example plot
Interpretation
6.2 Summary of Strengths
Caterpillar plots are good if you want to look at effects and variation at a glance to identify the best/worst (the extremes) or see the overall range.
Waterfall plots are good if you have a baseline to compare to for (1) a single time point, (2) many samples/patients, and (3) clearly know what above or below that baseline means.