CoChat

Lasagna Plots: A Saucy Alternative to Spaghetti Plots

Open access
Swihart, Bruce J., Caffo, Brian, James, Bryan D., Strand, Matthew, Schwartz, Brian S., Punjabi, Naresh M. · 2010 · Epidemiology
Cited by 107

The DOI matches its Crossref record; 3 of the 3 references it cites were found in a registry or an index.

Abstract

Longitudinal data are key to causal inference in epidemiologic studies. As the number of subjects, frequency of measurements, and period of active data collection grow, so does the size of the dataset. As the dataset grows, so do the burden and complexity of graphically exploring and summarizing the data without obscuring salient features. The gold standard for graphically displaying longitudinal data is the classic spaghetti plot, which plots a subject's values for the repeated outcome measure (vertical axis) versus time (horizontal axis) and connects the dots chronologically, using lines of uniform color and intensity for each subject. However, the classic spaghetti plot presents obstacles to the display of longitudinal data. Although useful for fewer subjects, trends and patterns are obscured with the larger numbers of subjects typical of modern epidemiologic studies. For example, trajectories commonly overlap in a classic spaghetti plot, as both subjects and the magnitude of the outcome measure are displayed on the vertical axis. With large datasets, the figure often succumbs to “over-plotting,” in which the multiple intersecting lines have no discernible patterns. One solution has been to plot a subset of the data based on medians or deciles, but this approach fails to use the whole dataset.1 Furthermore, repeated-measures data containing different enrollment times, missing data, or loss to follow-up (censoring), typically are difficult to display in the classic spaghetti plot. Exploratory plots generally are used to reveal various structures in the data: trends (how most people respond over time), outliers (whether some subjects differ from most), clusters (groups of patients responding the same, maybe due to a covariate such as treatment assignment), as well as data-quality checks of reasonable values and sensible collection patterns of multicenter repeated-measures data. Classic spaghetti plots suffering from over-plotting reduce the chances of seeing such patterns. Spaghetti plots can be improved. Color saturation, a method of handling over-plotting, has been implemented in parallel coordinate plots, of which the spaghetti plot is a special case.2–4 In many situations, a classic spaghetti plot or the more modern parallel co-ordinates plot can visualize data sufficiently. However, 2 common data features limit the usefulness of parallel coordinate/spaghetti plots: categorical outcomes and missing outcome values (eFigure 16 in eAppendix, https://links.lww.com/EDE/A401). We propose a procedure for visualization and exploration that is particularly useful for categorical outcomes and missingness. So-called heatmaps5 provide an alternative, or at least complementary, graphical-data-exploration technique. Heatmaps enjoy frequent use in the genomics literature and other areas with high-throughput data.6–8 However, in more standard longitudinal studies, they are less popular, as evidenced by the recommendation of spaghetti plots of the raw data1,9,10 or summarized data11 in popular texts of longitudinal data analysis. A lasagna plot is a heatmap well-suited for longitudinal data. In spaghetti plots, each subject's trajectory over time is like a noodle, that can cross other trajectories (Fig. 1). In lasagna plots, each subject's trajectory over time is a horizontal layer, with the simultaneous plotting of trajectories resulting in a stacking of layers, as in lasagna.FIGURE 1.: Lasagna plots as derived from spaghetti plots involve making noodles into layers. From left to right, (A) a spaghetti plot with 3 noodles where trajectories overlap. B, Extracting each noodle representing repeated measures on a subject, (C) a layer is made by letting color represent the outcome. D, Individual layers are then stacked to make a lasagna plot, with no overlapping of subject information.The proposed lasagna plot uses color or shading to depict the magnitude of the outcome measurement and fixes the vertical dimension per subject. The lasagna plot takes advantage of color to provide a third dimension and display additional information, rather than relying upon the vertical dimension to display overlapping magnitudes of change. All information about the value of the outcome is through color (intensity), making color choice important. Haphazard color selection can produce varying appearance in different media, induce optical illusions and after-image effects, and produce misleading interpretations of data characteristics via relative aspects of color.12–15 One good principle for color selection is to have equally spaced hues with constant chroma and luminance across hues in a perceptually uniform colorspace, such as hue-chroma-luminance (HCL) (as opposed to red-green-blue [RGB] or hue-saturation-value [HSV]).13 For unordered categorical outcomes, such as group membership, fixed chroma, and luminance for hues equally spaced within a hue-chroma-luminance colorspace provide a flexible framework for the generation of qualitative pa

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1097/ede.0b013e3181e5b06a

Cites

3 of 3 references verified0 citation errors · 0 not checkable

Statistics

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 3 references

Cited by

Metadata

Authors
Swihart, B. J., Caffo, B., James, B. D., Strand, M., Schwartz, B. S., & Punjabi, N. M.
Venue
Epidemiology
Published
2010-09
Volume
21
Issue
5
Pages
621-625
ISSN
1044-3983 (ISSN-L)
OpenAlex
W2094054668
Cited by
107
References
3
FWCI
7.00
Type
journal article
Language
en
Open access
bronze

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.