Points of View: Bar charts and box plots

Journal name:
Nature Methods
Volume:
11,
Page:
117
Year published:
DOI:
doi:10.1038/nmeth.2807
Published online

Creating a simple yet effective plot requires an understanding of data and tasks.

At a glance

Figures

  1. Variants of bar charts and a pie chart encoding the same data.
    Figure 1: Variants of bar charts and a pie chart encoding the same data.

    (a) Values in different categories are difficult to compare in pie charts. (b) Stacked bar charts enable comparison of overall values across items. (c) Layered bar charts support comparison of values within categories. (d) Grouped bar charts allow comparison of values across categories.

  2. Representation of four distributions with bar charts and box plots.
    Figure 2: Representation of four distributions with bar charts and box plots.

    (a) Bar chart showing sample means (n = 1,000) with standard-deviation error bars. (b) Box plot (n = 1,000) with whiskers extending to ±1.5 × IQR. (c) Probability density functions of the distributions in a and b. λ, rate; μ, mean; σ, standard deviation.

References

  1. Gratzl, S., Lex, A., Gehlenborg, N., Pfister, H.P. & Streit, M. IEEE Trans. Vis. Comput. Graph. 19, 22772286 (2013).
  2. McGill, R., Tukey, J.W. & Larsen, W.A. Am. Stat. 32, 1216 (1978).
  3. Krzywinski, M. & Altman, N.S. Nat. Methods 11, 119120 (2014).
  4. Spitzer, M., Wildenhain, J., Rappsilber, J. & Tyers, M. Nat. Methods 11, 121122 (2014).
  5. Krzywinski, M. Nat. Methods 10, 183 (2013).

Download references

Author information

Affiliations

  1. Marc Streit is an assistant professor of computer science at Johannes Kepler University Linz.

  2. Nils Gehlenborg is a research associate at Harvard Medical School and the Broad Institute of MIT and Harvard.

Competing financial interests

The authors declare no competing financial interests.

Author details

Additional data