Combining UX principles with Data Analysis
How to perfect your data visualization with simple UX principles
Friedrich von Schlegel’s “Every art should become science and every science should become art,” might sound a little over-exaggerated, but he does have a point — combining artistic and scientific approaches can only lead to something exceptional! When talking about data analysis the first things that come to mind are everyone’s not-so-favorite numbers and of course loads of information. This means that an analyst could have the most amazing findings present in the report and all his client would see are some random percentages and meaningless digits. This is where a creative approach takes its place in data analysis. In this blog, we’ll try to combine UX principles with data analysis to try and perfect our data visualisation processes.
What are data visualization and UX principles to begin with?
In one of the previous blogs we’ve guided you through the process of analysis, and the very next step is presenting your results to a client. For this particular part, one of the most important things to know is that in human cognition, understanding and memorability are closely connected.
Now you might ask “What will make my visualization memorable and hence understandable?” and we are here to help you. You’ve probably heard about UI/UX designers, but do you really understand what ‘UX’ stands for? According to Nielsen Norman Group:
UX (user experience) encompasses all aspects of the end-user’s interaction with the company, its services, and its products.
In the following sections, we will guide you through 3 basic UX principles to help you shape your visualizations into something truly remarkable.
Setting up the scene
Let’s imagine our client is a coffee shop owner who chooses music for their locale really carefully. The thing is, they don’t exactly want to play top Spotify charts since it’s usually mainstream and people get bored by repetitiveness, but they do want to play something people will find likeable.
Now that we’ve set the scene, we can move on to the UX principles and their connection to data visualization so we can help our client get a perfectly likeable playlist.
1. Place the User in the center
The first thing you want to keep in mind is who your client is and what their tech background consists of. Since you’ve been hired to analyze data for them, it’s very likely your client isn’t someone who fully understands how to get the needed results, and even if they do, they usually care more about the end product than the analysis process itself.
Considering what’s just been said, you should always aim to keep your results relevant and usable. This means you should exclude all parts of the analysis that don’t carry useful information for your client (Yes, even if it took you hours!). This principle can be applied regardless of the presentation way — writing a report or pitching a presentation.
For our example, presenting artists with most tracks in the top 200 chart (picture below) is only telling us about the artist’s current success. However, our client would like to choose music according to specific filters since top trending artists are constantly fluctuating.
Luckily, Spotify tracks data which includes parameters like danceability, tempo, genre and energy.
Using these attributes, we can present much more useful information which explains patterns present in tracks with most streams in the top 200 chart.
Now, these graphs are telling us something more exact — currently trending songs usually belong to the latin or pop genre. Additionally, most of these songs have danceability and energy above 0.5. With these parameters and sites for filtering Spotify playlists, our client can easily choose a specific subset of music with traits similar to the top 200 chart, and create playlists of likeable but not so mainstream music.
2. The Magical Number 7, plus or minus 2
Don’t let this subtitle confuse you; it is as simple as it sounds. This theory revolves around the idea George A. Miller proposed about the cognitive load of a human. All it means is that most adults can store between 5 and 9 items in their short-term memory.
Now, why is this important to us? When presenting results, you want to keep this theory in mind and remove all unnecessary elements from the visualizations. For example, if you want to show the frequency distribution of a single field, don’t present all occurrences, instead reduce the number so people watching are not overwhelmed and can process the values.
In our example, we can only focus on the top 7 artists with the highest number of streams on the Top 200 chart.
3. Provoke emotion and delight
It has been proven that beautiful things evoke positive emotions. In fact, the aesthetic-usability effect says that when people find a design visually appealing, they may be more forgiving of minor mistakes. Use this principle to your advantage, but be careful to keep the focus on the information you want to present and not the design itself.
One of the few things you want to start with is choosing the right colors. In most visualization tools, you are offered default colors, which can sometimes be too intense for the viewers. Reducing the saturation, or intensity, of colors on a view screen usually helps with this problem.
Additionally, if the data in a chart is differentiated into one bar per artist, there’s no need to differentiate it with colors as well. Instead, choose one color for all bars and help your viewers focus on all values in a bar chart equally.
Results of updating our previous chart according to the mentioned rule:
For the opposite effect, try bringing attention to a dimension member, by using a bright color, and keeping the remaining bars in a neutral one.
Extra tip: Use associations with colors for a bonus effect. For example, when pointing out that something represents an issue use red and for desirable outcomes use green. Also, avoid using these two colors when presenting neutral results.
Final words
Sometimes, it is not easy to represent all of the findings from your thorough analysis in a short report, but it is your job as an analyst to differentiate the important results from the less important ones. What should also be part of your decision-making process is using tips like the ones mentioned above when trying to make a statement and a visually appealing presentation.
Now that you’ve learned some science behind popular and likable music, try to modify and filter out your favorite playlists and let us know how many songs would be suitable for our fictional coffee shop!
Tech Bite by Lejla Vardo, Data Analyst at Atlantbh.
Originally published at https://www.atlantbh.com on June 10, 2022.
