Visualization speaks more than a thousand numerals

Visualization of the different layers of matter in a neutron star. The red core consists of quark-matter. Annala E., Gorda T., Kurkela A., Nättilä J., Vuorinen A. Evidence for quark-matter cores in massive neutron stars. Nature Physics, Letter 01 June 2020.

Visualization speaks more than a thousand numerals

Visualization plays an important role in science. Visualization brings research to life. Scientific visualizations are nowadays invariably used when communicating science to the general public, for example in television news

From numbers to images

In the first decades of computers, they were only used to make calculations. They produced and printed numbers. Gradually, many technical-visual industries such as the film industry, the gaming industry, and the advertising industry realized the potential of computers in image making. The development of computer graphics was driven in part by the popularity of GUI based operating systems.

In the wake of rapidly evolving computer graphics, the U.S. National Science Foundation (NSF) commissioned a study whose 1987 report, “Visualization in Scientific Computing,” highlighted the importance of visual illustration of scientific results in international technical-scientific competition.

In the same year, 1987, CSC acquired their first graphics workstation. In the early 1990s, a visualization laboratory and a scientific visualization program developed by CSC were made available to researchers.

Visualization had become a new scientific method also in Finland, which allowed researchers to better view and understand the growing amount of data. In ten years, equipment and software became cheaper to a level where individual research groups could acquire them for their own use.

 

Visualization of an atmospheric airflow, introducing several visualization techniques. Simulation Cassiani M., Ardeshiri H., NILU - Norwegian Institute for Air Research.

 

 

Making science visible

 

Visual imagery conveys the principles of research to the viewer, and the experience can guide behavior, as animations about the spread of coronavirus have shown.

Due to corona pandemic in the spring of 2020, the importance of scientific research and presenting the results visually increased. The virus cannot be seen, nor can its spread be understood in the same way as, for example, the smoke and smell of tobacco. Because it cannot be smelled, tasted, heard, or seen, it must be made visible by using visualization.

CSC’s scientific visualization expert Jyrki Hokkanen helps researchers to illustrate scientific data. He also maintains visualization software and assists in its use. All the visualizations on this article are his work.

- Calculations and measurements produce numbers. People have evolved to interpret things they sense in their surroundings, not series of numbers. The brain is at its most accurate when analyzing images, so it is natural to visualize the research results into images, Hokkanen says.

However, the presentation must be made as demonstrative as possible. In addition to visualization, hearing is quite useful for various observations, for example in a case of metal detectors.

- The easiest way to illustrate radioactivity, for example, is the clicking noise of a Geiger counter, which frees the sight for other uses.

 

Visualization of droplet sedimentation in turbulent airflow inside a room. The small blue particles remain suspended in the air like smoke, while the heavier red droplets fall slowly to the ground. Vuorinen V. et al. Modelling aerosol transport and virus exposure with numerical simulations in relation to SARS-CoV-2 transmission by inhalation indoors. Safety Science, Volume 130, 2020, 104866.

 

 

Information can be presented in different ways

 

Scientific visualization illustrates phenomena with a natural spatial form, such as galaxies or cracks in an iceberg, whereas information visualization describes abstract statistical measures such as the development of stock prices or website visits.

- Terminology easily leads to misconceptions, because in information visualization the data can be scientific. Both methods may be used to illustrate data from the same subject area, Hokkanen emphasizes.

In scientific visualization data points are usually represented as surface or volume elements, or as marker symbols called glyphs. In information visualization the main challenge is often in choosing the appropriate statistical methods. The end results are depicted as graphs and maps.

Visualization is not always just the last step in the research process, as intermediate results are also visualized. For example, visual inspection during the computations makes it possible to adjust the parameters and steer the simulation in real time, so that the model is made to describe the subject in more detail.

Scientific fields that use a lot of scientific visualization include molecular research, medical imaging, biology and many areas of physics such as mechanics and astronomy.  Scientific visualizations most familiar to public include atmospheric simulation results, i.e., weather reports.

Visualization of glacier ice calving, with a ship added for 100-meter scale. Simulation done at CSC. Åström J., Vallot D., Schäfer M. et al. Termini of calving glaciers as self-organized critical systems. Nature Geosci 7, 874–878 (2014).

 

Difficulty of visualization

Visualization is often done with the same computer program with which the data was produced. In more challenging cases, this is not possible, or at least not up to the desired quality, which requires special visualization programs. Transferring large amounts of data between computers can take days, so the aim is to produce the visualization where the data is located. However, sometimes this is not possible.

Once the data is within the reach of the visualization program, it is modified into a format that the program understands. Rewriting large data sets can be slow. Once the data is in the correct format and transferred into the visualization program, it is viewed interactively from different angles. Data can be sliced or pieces cut off to gain insight of the meaning of the variables.

When rotating and translating data, the computer may have to render images whilst it is operating at the limits of its capacity. In these situations, the display becomes jerky and lagged as the thousands of processors on the video card cannot perform their calculations in real time.

Visual impact or scientific rigor?

The cover image of a scientific journal has a different function than the images in the research papers inside a journal. The cover is designed to draw attention and does not need to provide accurate research information. On the other hand, the visual style of research papers is expected to have a dry sense of purpose and seriousness.

- Researchers also have to promote their research work. Impressive images are an important part of the visibility of the research and can have an impact on funding decisions, says Hokkanen.

Read more on CSC's Visualization Services

CSC grants access to several applications with sophisticated visualization capabilities, such as VMD and pyMOL for molecule chemistry. CSC also maintains two general visualization applications, ParaView and VisIT that are Open Source and available from the manufacturers website for local install onto a desktop, or onto a virtual computer such as cPouta.

More about this topic » Go to insights and news »

Hannu Mourujärvi