The Role of the User Interface in Data Visualisation
Many consider a walk in the park a relaxing experience. The air feels and smells pure, and our mind is calm in the simplicity and safety of putting one foot in front of the other. Our brain is relaxed because it doesn’t really have to think, but to our subconscious mind, this task is not as simple as it may seem. Our mind receives millions of discrete packets of electrical information every second from our senses. Just looking at a leaf involves 200,000 times more information than what our conscious mind can handle. That is so much information that we are constantly processing and converting this information into simpler perceptions for our conscious mind. We are so good at processing and presenting that our perception doesn’t even hint at the tremendous effort required to produce the final image we see or smell we feel. This way, we can enjoy our walk in the park without being overwhelmed by information.
Modern designers face the same challenge today. How does one take so much information and condense it into the most easily understandable form possible? No one wants to overwhelm their user. In fact, overwhelming data may be the very reason the user came to you.
Inrobin’s users want data clarity, and often, big decisions are at stake. When faced with this problem, we have identified 3 points that make users’ data lives easier
When you have more data than you know what to do with, it is especially important to focus your offering. Not because you can’t analyse everything you have, but because your user will be confused by and doesn’t need to see everything you have. How can you do this? Ask the user what challenges they have, what takes away their time, what they can’t function without. And find the right user, if possible. Sometimes, your contact will say a lot but won’t use your product at all. Third, focus as a company on finding the right problems to solve or outcomes to address. This may be the most difficult thing to define. However, be open to learn from others during your company, life, or academic journey and this information will materialise in time.
Knowing what overall problem to solve and what your user needs will make it easy to prioritise what kind of data you need to show. However, the magic bullet may be analysing the data to find which variables truly make a difference to move towards the outcome you are measuring. Find out the variable which gives data which most closely indicates the outcome. Do you want to know how to make an oven as hot as possible? Displaying the temperature of the oven will give you the best indication of your outcome. Do you know how temperature increases in your oven? Measuring electric current through the oven will give you another indication of how hot the oven may be. Inrobin uses Machine Learning in real time to determine which variables are most likely to indicate the outcomes we measure, and we display those variables on our interface. This eliminates much of the noise we could analyse and instead focuses the user’s attention on the most relevant details.
This is applicable to every product, but becomes less obvious with very complex offerings like large machines, software, and services. As the task becomes more and more challenging and abstract, it gets easier to let the difficulty of overcoming the design tasks compensate for a product that works every time. Although one may not be able to offer a 100% reliable product, it is important to keep reliability at the top of the list of priorities. Design and data presentation become irrelevant if the data presented does not provide correct information, or at least is not completely honest about its capabilities. What this means is, use wording that makes it clear how reliable information is (e.g. stating the probability of some statement or giving intervals of confidence for predictions) and use visualisations that explicitly avoid deception. One classic example of deception is representing two similarly-priced products with such a small range of values on the axis that the price of the competitor’s product looks drastically higher than that of the product being offered.
Ultimately, the benefit of being reliable is that it allows the design to impact the user. Without reliability, design is just decoration and misleading statements.
3. Taking Advantage of Human Processing
Essentially, there are two things to take advantage of human visual processing. First, just like the laws of physics, everything should be ordered or follow a consistent rule. The first thing shown should be the most important, and the last thing the least important. Or the most recent thing is first and the oldest thing is last. Or the biggest image represents the most important thing and the smallest thing is the least important. This is most important when organising different types of visual information or when including a table. That way, the user knows that, for example, the first graph to show up is the most important graph, or that the data at the top of a table is the most recent data. Second, movement and changes can greatly increase the understanding of an image when used with purpose. The movement of an object describes almost as much about an object as its image, and it is easy to see why when you contrast a picture of a human-like robot (fairly similar to a real human) to a film of a human-like robot (showing its cold, unresponsive movements compared to a real human’s movements).
This is the basis for the construction of all the types of graphs. It is also the basis upon which one can make their own creative graph types which are more informative but even easier to understand. I will focus the rest of this section on graphs because they are designed to take advantage of our natural processing in the simplest way possible. Even then, they are not intuitive enough for most users. To avoid this pitfall, a designer must become smart about graph formatting, for which there are many excellent articles on the internet (e.g., Rules for Graphs, Rules for Interactive Graphs).
Comparisons are easy with bar charts. If the values possess a percentage of the same resource, a pie chart is better.
Status of a variable can be represented as a sliding indicator between an average minimum and an average maximum value, such as a speedometer in a car or the level of an analog thermometer. This type of display gives not only a real-time value, but also context. Context is invaluable when you want to describe normal machine function to a new user.
The heat map is good for representing a value that changes over a surface, such as temperatures on the surface of the Earth.
The famous line graph is excellent for comparing values over time. One can also use it to see how a variable changes based on the value of other measurements, such as wind speed at different elevations. The most information-heavy graph,and, therefore, the hardest one to present well, is the
multiple-line graph. In this case, you are comparing several variables over time or over another measurement. It is also possible to compare slopes, line behaviours, and cause and effect between lines which may not measure the same variables. The critical task is to include a focus for the multiple-line graph. This could be with the title, where the focus of the graph is clearly stated, e.g., “How Changes in Acceleration Affect Velocity” or “Changes In Variables Over Time”.
Lastly, there is the possibility of making interactive graphs. This may be the most useful feature of modern data visualisation, whereby one can change and edit a graph as a user interacts with the graph via clicks, touches, or hovers. It takes advantage of various aspects of human processing.
• First, it cleans up the appearance of the graph, because not everything has to appear on the screen at once, and the interactive graph program can show very specific and relevant data given the right context (revealed by the position of the mouse, or finger). For the same reason, the graph can contain more information than a classic graph, and can provide information only when it is necessary.
• Second, the interactive graph takes advantage of changing states. These are very noticeable movements. For example, a mouse pointer hovering over a line can cause all other lines to fade away into near transparency while the line of interest remains completely opaque. Updates can happen in real time and the flow of data is noticeable due to the developing points at the edge of the graph.
• In conclusion, the interactive graph provides ample customisation for the user and for different information states. Whereas a designer must guess at the most useful information for a user in an average situation, an interactive graph allows the user to be their own designer for different situations, and allows the programming to adjust the design for emergency situations or for contractors and different types of users. This allows the designer to make useful changes that enhance the user’s experience when the user needs different types of information.
• If interactive graphs have one downside, it is the time it takes to create one. A static image can be easier to develop than an interactive one. With the tools available today and in the near future, the difference in difficulty between creating interactive and static graphs may become insignificant.
The amount of data society collects increases every day with new data-powered services like Siri and Google. It is possible, with the graphing technologies available today, to incorporate interactivity features with time-tested graphical visualisations to create simple, information-rich representations of all the information passing through our databases. Inrobin offers solutions to make sense of industrial data for their clients to reduce risk of their machinery, and there are many opportunities to use the same data to create even more value. Our senses collect new data every day, and even old data can often surprise us with new insights when revisited. Through our efficient processing system, our mind is often not overwhelmed, and perceiving new things and places can even be pleasurable. Therefore, we keep going back to the park, time and time again, to relax our minds amid the barrage of data.
By André Faroni, Electromechanical Engineer, Inrobin Ltd.