2.3.4 Data visualization

Jump to: navigation, search

​​​​​​A. Background & Definitions

Data visualization serves to enhance readers' understanding of the data and to provide support for appraisal of the data. In the scientific ​​publications, authors frequently use the static graphical forms to show the data that support key findings (Weissgerber et al. 2015). Technically more challenging as it requires specific tools and skills, interactive graphics may improve data presentation in scientific publications by allowing readers viewing different types of graphs plotted based on the same data sets, improve readers' understanding of the data and allowing for adequate evaluation of data (Weissgerber et al. 2016​ and Weissgerber et al. 2017).

It should be remembered ​​that many of the expectations about the best practices of graphical representation of the data do apply to tabular and other forms as well.

B. Guidance & Expectations

Dat​a analysis should be performed exactly as described in the study protocol and figures, which represent data and should provide sufficient information allowing to understand and rigorously interpret data (the following recommendations are based on Motulsky 2015​, Weissgerber et al. 2015):

  • For scientific publications, it is recommended to use figures that show the distribution of continuous data, i.e. univariate scatterplots, box plots, or histograms.
  • For studies with small sample sizes, it is recommended to use univariate scatterplots, which is the most informative graphical presentation.
  • Bar and line graphs are not recommended for the use in studies with small sample size. Such graphs do not show data distribution as the same bar or line graph can be plotted based on substantially different datasets.
  • If data distribution cannot be presented, it is recommended to plot mean with error bars that represent the standard deviations (mean ± SD), or means with error bars showing the 95 % confidence interval (CI) of the mean.
  • Researchers are encouraged to provide information if and where raw data are available.

As the ​data visualization cannot be separated from data analysis, it is strongly recommended to report all details when describing data processing and statistical methods (Motulsky 2015):

  • it must be clear what experimental unit is (e.g. biological replicates, technical replicates​​, etc.);
  • if any observations were excluded, it must be clearly stated how many were eliminated, the rule used to identify them, and a statement whether this rule was chosen before collecting data;​
  • if data were subjected to normalization, please explain exactly why and how you defined 100 and 0 %;
  • explain the details of the statistical methods you used. For example, if you fit a curve using nonlinear regression, explain precisely which model you fit to the data and whether (and how) data were weighted.

Finally, researchers are strongly encouraged to visualize the data in a way that supports reporting of the effect sizes with expressions of uncertainty (Calin-Jageman and Cumming 2019).​


  • To consider adding this subject to a training program for new employees or refresher training (if appropriate)
  • To apply the same principles and standards to graphical and non-graphical forms of data visualization


C​. Resources

  • Motulsky HJ (2015) Common misconceptions about data analysis and statistics. Pharmacol Res Perspect. 3(1). [1]
  • ​Weissgerber TL, Milic NM, Winham SJ, Garovic VD (2015) Beyond Bar and Line Graphs: Time for a New Data Presentation Paradigm. PLoS Biol 13(4): e1002128. [2]
  • Weissgerber TL Savic M, Winham SJ et al. (2017) Data visualization, bar naked: A free tool for creating interactive graphics. J Biol Chem. 292(50):20592-20598. [3]
  • Weissgerber TL, Garovic VD, Savic M et al, (2016) From Static to Interactive: Transforming Data Visualization to Improve Transparency. PLoS Biol. 14(6). [4]
  • Ellis DA, Merdian HL (2015) Thinking Outside the Box: Developing Dynamic Data Visualizations for Psychology with Shiny. Front Psychol. 6:1782. [5]
  • Calin-Jageman RJ, Cumming G (2019) Estimation for Better Inference in Neuroscience. eNeuro 1 August 2019, 6 (4) ENEURO.0205-19.2019. [6]

Journal guidelines:

  • ​PLOS Biology (2016) Submission guidelines: data presentation in graphs [7]
  • ASPET journals' instructions for authors (detailed explanations with examples) [8]

Online tools for creating interactive graphics:

  • Allows to create interactive line graphs [9]
  • Allows to create customized interactive graphics such as univariate scatterplots, box plots, violin plots and, only for educational purposes, bar graphs [10]
  • To create interactive graphics from data obtained with repeated independent experiments - [11]

back to Toolbox

Next item: 2.4 Reporting