In this post, we are focusing on data analysis and visualizations, which do not come from the devil, as many leaders of organizations think. We believe that data analysis and visualization can be understandable, discoverable, and manageable for the average person.
Why Do People Not Trust Computer Analyses and Visualizations?
The question is not that simple. People overrule, over think, and misunderstand the computed data while they commit their lives to a medical device without question (for example, the values of an EKG)
What Stands Behind This Distrust?
In our opinion, old, outdated, ugly, and useless software plays a big role in distrust. In this type of software, the user interface is made for professionals (mathematicians) and seem useless to the average person. This perception causes many thoughts such as, “I am not skilled enough for it,” or “This is not my job.” On the contrary, the analysis process has to be done by the owner of the data, who knows and understands it.
If someone else does the analysis, (for example the analytical department) the results could be different or irrelevant.
This trend is changing nowadays, and the number of new, modern visualization tools on the market is increasing, but people rarely change, and they tend to rely on their bad experiences.
In the following sections we will describe some typical problems and suggest solutions.
When getting a chart or data point that proves a certain idea, most people think the problem or notion is already solved.
Company A has suppliers. The supplier/delay chart shows that the deliveries from the biggest suppliers are regularly delayed. Therefore, Company A believes that this delay is the reason why it cannot sell more products. However, if we approach it from a different viewpoint (for example, the product/trend chart), we can see that the product has less demand than it did previously.
It is worth exploring our data in different views or globally. If we investigate it in just one way, we cannot be sure that we will obtain the correct result we just see it one perspective. The relations in the background also matter.
Many people are not able to abstract from their thoughts. If someone says something different (in this case, the computer), they believe it is definitely wrong.
Over the last couple of months, the demand for Product A has shown a growth trend. This month, it decreased, but the market has not changed at all. We decide that this decrease is impossible, that we measured poorly, or that something is wrong. However, it is possible that the growth trend was caused by some kind of external factor and not the result of our work.
We should try to verify our data with facts, and relations. We must not rely on only our intuition but rather be open-minded
If we scan our data in many aspects, we might find cases that are not closely connected to our task or problem.
We want to see our income from Product A in segments. If we have too many segmentations, there will be a significant outlier sooner or later caused by a random fact.
We should focus on the task or problem to be solved, investigate the circumstances and what causes and effects can be possible between them, and filter out the outliers.
If we tried some data analysis and visualization software before and found it useless, we cannot be sure that there is no right software on the market for us.
We once tried a data analysis software, but we could not load our data because the UI was too confusing.
We should try another data analysis software because it may have very different features, and we might find the right one for us.
Everyone should use data analytics and data visualization during his or her work.
Nowadays, everything is turned into data. Almost everything is connected to one another.
Data mining and data digitalization are much more easily achieved nowadays. We can definitely say that our organization has to be data-driven because there are many ways to optimize our success or increase our income. Be data driven!