JHERONIMUS ACADEMY OF DATA SCIENCE
  • Home
  • About us
    • Vacancies
    • Convent
    • Campus
  • Education
    • Bsc Data Science
    • MSc Data Science Entrepreneurship >
      • Curriculum
      • Study program
      • Admission
      • Pre-Master's Program
      • Scholarships
      • Info activities >
        • Msc DS&E event
      • Career in Data Science
      • Campuslife at JADS
      • Student Housing in 's Hertogenbosch
      • Rules & FAQ
      • Examination Board
    • All Master's Programs Data Science
    • PDEng Data Science >
      • Application PDEng Data Science
    • Professional Education >
      • Discover Track
      • Execute Track
      • Data Science & Health
      • Data Science & Public Services
      • Data Science & Industry
      • Data Science & Energy
      • Data Science & Professional Services
      • Data Science & Supply Chain Management
      • Lead Track
  • Research
    • DSC Den Bosch >
      • Research Projects
    • DSC Eindhoven
    • DSC Tilburg
    • Research showcases
  • Business
    • Partnerships
    • JADS Playground
    • MKB DataLab
  • News
    • Newsflash
  • Contact
  • Privacy statement

12 April 2017
​
Learnings from JADS Masterclass on Data Visualization 


Text: Christian Markwat

Jack van Wijk starts the Masterclass with a question to the audience: Why use data visualization and what is it good for? Most participants do know why and actively contribute by answering his question. This very question (and its answers) appeared to become a red line during the Masterclass. We all think we know a lot on how to visualize data properly, but there is much more to it than just creating a 3D Pie Chart in Excel! 

The collection of a larger or smaller amounts of data often starts in the form of tables. Tables can be handy to look up and compare values, but it does not give a quick overview of how the data points are related to each other. Moreover, tables may become unreadable when large amounts of data points are included. To get more insight in the numbers, data visualization is a good idea!
Why visualize?
Why should we use pictures (visualizations) to show data? A key reason is that visualizations can show outliers and detect numbers that are ‘invisible’ in tables. Next to that, visualizations can compress similar data points and easily show their impact on the relationships that you are looking at. Data visualization can also be dangerous as it is not always presented in a trustworthy manner, for instance when making very small differences look very big (e.g. the Y-axis does not start at 0). 

" Data visualization can be dangerous as it is not always
​                                    presented to the user in a trustworthy manner "


During the Masterclass Van Wijk showed examples of data visualizations that showed a completely different picture than the data told. Some examples include visualizations presented to the public that are complex to the extent that nobody actually understands them. Van Wijk: “Sometimes data visualizations look like pieces of art, but they do not tell a lot…”.

​
Contemporary scholar with practical insights
To better understand how to present data, Van Wijk provided four example of historical data visualizations. These scholars were quite ahead of their time in presenting data in a very understandable manner and are worth a digital look:
  • William Playfair (1786) on trade between Denmark and Norway
  • John Snow (1854) on the origin and spread of cholera in the city of London
  • Charles Joseph Minard (1869) on Napoleon’s campaign against Russia
  • Francis Anscombe’s quartet (1973) shows four data sets with equal simple statistics, but dissimilar graphic representations.

An interesting example of more contemporary data visualization is the 2006 TEDtalk by the late Hans Rosling (link), co-founder of the Gapminder Foundation. He shows statistical ‘myths’ about the developing world. Anyone who is interested in the impact of data visualization is definitely recommended to view the video.

Participants of the Masterclass are very enthusiastic about the methods of data visualization that were presented. During the break, several participants discuss that these visualizations: "(…) it bring the data alive…" and "(…) it gives context to the data".
Picture
Prof. Jack van Wijk
​Jack van Wijk is a full professor in the field of Visualization at the Department of Mathematics and Computer Science of Eindhoven University of Technology. He received a MSc degree in Industrial Design Engineering (1982) and a PhD degree in Computer Science (1986), both from Delft University of Technology (with honors). After a short period in the software industry, he has worked for ten years at the Netherlands Energy Research Foundation ECN. He joined the University of Technology in Eindhoven in 1998, where he became a full professor of Visualization in 2001. His main research interests are information visualization, visual analytics and mathematical visualization. 
​

> Read more about Jack van Wijk
Business graphics
After a light meal during the break, Van Wijk continues with the subject of business graphics, including business oriented data visualization methods. He explains the visualization pipeline. Which is the (step-wise) process of creating visual representations of data. This process starts with the raw data. The first step is to clean up the data, this is called transformation of the data. Then the data needs to be mapped (e.g. in geometric objects) and projected (e.g. as an image). The final step is to present the data to the user.
 
In this part of the Masterclass, Van Wijk presents several methods for Information Visualization and Scientific Visualization, each with its own focus. A plethora of possibilities exist when it comes to visualizing data. The website Visual Complexity is a top tip if you are interested in new data visualization methods, it covers hundreds of interesting approaches.
 
The use of symbols and color palettes need to be carefully considered too. Also, the size of the symbols or lines in the graphic matter, if only to increase readability of the graph. For help with color palettes have a look at the excellent website of ColorBrewer by Cynthia Brewer. Here you can construct a custom color palet based on your choice of number of data classes. There are even options to create color palettes that are color blind safe, print friendly and photocopy safe. 
 
Van Wijk presents five tips for presenting understandable business graphics to a broader business-oriented audience:
  • Practice making business graphic
  • Be critical about readability of the graph
  • Do not just use default settings
  • Look around for good (and bad) examples
  • Try to analyze and understand why things do or do not work
Picture
Visual Complexity 
"Functional visualizations are more than innovative statistical analyses and computational algorithms. They must make sense to the user and require a visual language system that uses colour, shape, line, hierarchy and composition to communicate clearly and appropriately, much like the alphabetic and character-based languages used worldwide between humans." 
​
Matt Woolman - Digital Information Graphics


> View more examples
Challenges in Data Visualization
In the remainder of the workshop, Van Wijk addresses two challenges in data visualization, especially when presenting large amounts of data and multiple variables: (1) How to handle big, messy data? and (2) How to deal with models.
 
Standard diagrams do not scale enough when using a lot of numbers or variables. E.g. scatterplots become unreadable. Focus on visual representation, without communicating too much variables. The use of computer-supported interactive visual representations is becoming more and more common.
In the real world, data is often messy. It can be incomplete, include a lot of noise and it is heterogeneous. Abstract data is often a mix of scatterplots, tree diagrams and node-link diagrams. For instance, software packages are connected in many ways. Hierarchy plays a more important role and it helps to bundle data.
 
In the final part of the Masterclass, Van Wijk presents some of his own research in data science and data visualization. His SequoiaView diagrams were particularly interesting to see as they make use of completely different methods of presenting data. Have a look at the website of SynerScope for more information and inspiration on this topic. Read more about SequoiaView here.
Upcoming Masterclasses
Interested in learning more about these topics through the story of experts and explained in a practical way? You are most welcome to join the upcoming Masterclasses!

Next up: Beyond privacy and compliance: data ethics as a practical tool for understanding risks and remedies, presented by
​dr. L.E.M. (Linnet) Taylor of Tilburg University.


> Read more and register

Conclusion: One Big Lesson
The big lesson of this Masterclass is that data visualization need to be carefully considered if you want to produce quality business graphics that can be presented to your target audience.

​Data visualization of big messy data is another story. Expert techniques are needed to visualize this data (e.g. SequoiaView). With an abundance of data available nowadays it is becoming more and more important to invest is skills to uncap the potential of this data and visualize it. 
​
​Recommended reading
  • Show Me the Numbers: Designing Tables and Graphs to Enlighten – Stephen Few
  • Now you see it: Simple Visualization Techniques for Quantitative Analysis – Stephen Few
  • Information Visualization - Perception for Design (3rd edition) – Colin Ware
  • Visualization Analysis & Design – Tamara Munzner
  • Weapons of Math Destruction – Cathy O’Neil

About JADS    
​The Jheronimus Academy of Data Science is an ambitious initiative of Tilburg University, Eindhoven University of Technology, the City of 's-Hertogenbosch and the Province of Noord-Brabant. The Academy offers seven bachelors and masters programs at three locations (TU Eindhoven, Tilburg University and Mariënburg Campus). Plans are to eventually be able to accommodate between 1,500 and 2,000 data science students from around the world. The programs are innovative and truly multi-disciplinary. This is especially true for the Mariënburg Campus where researchers and students work in close collaboration with industry within a unique educational concept.

> Contact us

Education
Research

Business


Privacy Statement
About us

Vacancies
News
​Contact


Foto
Disclaimer
  • Home
  • About us
    • Vacancies
    • Convent
    • Campus
  • Education
    • Bsc Data Science
    • MSc Data Science Entrepreneurship >
      • Curriculum
      • Study program
      • Admission
      • Pre-Master's Program
      • Scholarships
      • Info activities >
        • Msc DS&E event
      • Career in Data Science
      • Campuslife at JADS
      • Student Housing in 's Hertogenbosch
      • Rules & FAQ
      • Examination Board
    • All Master's Programs Data Science
    • PDEng Data Science >
      • Application PDEng Data Science
    • Professional Education >
      • Discover Track
      • Execute Track
      • Data Science & Health
      • Data Science & Public Services
      • Data Science & Industry
      • Data Science & Energy
      • Data Science & Professional Services
      • Data Science & Supply Chain Management
      • Lead Track
  • Research
    • DSC Den Bosch >
      • Research Projects
    • DSC Eindhoven
    • DSC Tilburg
    • Research showcases
  • Business
    • Partnerships
    • JADS Playground
    • MKB DataLab
  • News
    • Newsflash
  • Contact
  • Privacy statement