“In God we trust, all others bring data” , William Edwards Deming

In modern football behind every dominating team there is a mastermind tactician. All success stories involve world class managers who lead their teams into great victories. There are several examples including Jurgen Klopp^[1], Pep Guardiola^[2] or Unai Emery^[3]. All these “architects” prepare their teams, transform their players, and extract the best performance in the pitch. To achieve that, there are tiny details and huge decisions they must consider, making them stand out. Their job may seem ideal, but they are in a continuous evaluation and depended on their results. As William E. Deming^[4] quoted: “In God we trust, all others bring data”. Indeed, football managers must continuously make decisions that bring good records to feel safe in their position.

The rule of good performance and difficult decisions applies not only for sport clubs, but also for almost every organization. The good news is that today with the evolution of technology, the humanity has created many tools to minimize the risk and optimize the decisions and their outcomes. At this point we will focus on the contribution of descriptive analytics in a business and their essential use in decision making process. This part of business analytics is about describing past events, things that have happened and past facts that constitute our data. They need to be presented in such a way that the receivers should not be led into wrong conclusions or misunderstandings that would affect their effectiveness in decision making. From the descriptive analytics we are going to go deeper and isolate the data visualization area and the tactics that should be applied to achieve our purposes.

For data visualization it is fundamental to ensure that our graphs have integrity, present the things how they really are and accurately inform the receiver. According to Tufte^[5] there are 6 basic principles that should be respected to achieve rectitude in our visualizations:

• -Proportionality
• -Clarity
• -Standardization
• -Appropriate dimensions
• -Contextualization
• -Data driven

The aforementioned key elements contribute to the integrity of the visualizations which provide to the audience the tools to define and make the best decisions. Proportionality ensures that the illustration of numbers, as displayed in the graphic itself, should be directly proportional to the numerical quantities represented. With clarity the clear and detailed labeling is protected to avoid ambiguity. When we need to present time-series of money, standardized units are on most occasions better than nominal. It may seem obvious, but the number of our dimensions depicted should not exceed the number of dimensions in the data. This makes sense if we consider that we are displaying a graph with two dimensions but labeling three. This of course will create ambiguity. In addition to the previous element, contextualization pushes us not to mention in our graphs any data that is out of context. Finally, our visualizations should be data driven, which means that they should display any variation that exists in the data and not create any. In simple words, the audience should discover any real patterns and not get manipulated by the designer of the graph. After examining those tactics in our graphs, we are now ready to build reliable visualizations to boost the efforts of any decision-maker.

Galanis Dionysis, Professional Services BI Associate, WITSIDE