Flexible Trees

Flexible Trees

Flexible Trees allow for creating tree visualizations to fit a desired shape by using sketch-based interaction. It offers a simple way of generating tree layouts for infographic purposes by adding a meaningful shape to a tree visualization.


  • Adobe Illustrator
  • Adobe Photoshop

How It Started

The project was done at the InnoVis Group led by Sheelagh Carpendale at the University of Calgary in Canada. The project started, long before I visited Calgary, as a computer graphics algorithm to fit layouts of tree visualizations (mostly node-link diagrams) into a desired shape. The shape is directly drawn on top of the tree visualization, which leads to the tree being adjusted to the pen input. The original idea for the project was to create tree visualizations with meaningful shapes, that could serve as a foundation for infographics.

What I Contributed

For a final publication of the project, we needed compelling demonstrations for using custom shaped tree visualizations as infographics. This was the point that I joined Flexible Trees. I was asked to design and create infographics for the publication that could show the benefits and strengths of Flexible Trees. I searched for tree-based data and came up with three possibilities of using the shape of a tree to convey meaning.

Add Context

For the first use case, I tried to contextualize a tree visualization by creating the shape of an object that is related to the topic. I chose the shape of a coffee cup that hints at the topic of the tree within the cup: the various coffee aromas provided by the SCAA Flavor Wheel.

Infographic about coffee aromas

Add Quantitative Data

I also wanted to incorporate more data in the shape of the tree. After some testing, I decided on a pie charts. While definitely being controversial, they’re still one of the most popular visualizations. The pie chart shows the top causes of death for the United States, Germany, and India (Source: WorldLifeExpectancy). The shapes of the slices add further quantitative data: slice angles represent birth rates for each country. In addition to conveying the information contained in the trees, this infographic puts the information in perspective by also comparing the areas of the trees in the pie chart.

Infographic about death causes and birth rates of three different countries.

Add Context, Qualitative, And Quantitative Data

Combining all approaches led to trees whose color distribution and slice angles represent the impact of each cause of death on the total percentage of deaths. (Source: WorldLifeExpectancy) The shape, furthermore, adds information to the tree: it conveys the country that the data is related to. Filling parts of the countries also conveys quantities like a pie chart does.

What The Progress Was Like

After having found suitable data to turn into a tree visualization, I transformed the data into a format that the algorithm could read. I ran the algorithm with the data on a computer that supported pen input which I used to draw the outline for the tree adaption. The output of the algorithm consisted of an image that gave me the foundation for the infographics. Consequently, I imported the output images into Adobe Illustrator and used Illustrator as well as Adobe Photoshop to create the infographic based on the output image of the algorithm.

Sketching the outline of Germany for the algorithm to adjust

Lessons Learned And Further Information

Looking back at the infographics I made in 2016, I still like their visual design and the ideas that they convey. However, I’ve became a bit more selective and critical about the data sources. I think in retrospective, I would be more careful in choosing trustworthy data sources.
The project resulted in a publication that I presented at the AVI (Advanced Visual Interfaces) conference 2016 in Bari, Italy.

Javad Sadeghi, Charles Perin, Tamara Flemisch, Mark Hancock, and Sheelagh Carpendale. 2016. Flexible Trees: Sketching Tree Layouts. In Proceedings of the International Working Conference on Advanced Visual Interfaces (AVI ’16). Association for Computing Machinery, New York, NY, USA, 84–87. DOI: https://doi.org/10.1145/2909132.2909274