7. Data visualization

Summary

Within data analysis, the person who builds visualizations may hold a separate role. It is an area that involves technical knowledge not only of data but of design. Without this specific training, it is often acquired through experience, and even then, in a presentation, content and form will depend on who receives the information. The advisable approach is to start from a minimal, replicable design and take cues from other presentations, and then to complement and improve it as the basic principles of design, color and even trends are absorbed.

As for the data, the visualizations will be determined by the number of variables and their type. It is usually a good habit to consult a visualization gallery to check which chart fits what is available or what needs to be shown (for example, Python Graph Gallery). The visualization should always be adapted to the question it must answer, avoiding overload / repetition of data, and accompanied by the corresponding title, labels and legend and, where needed, a brief, plain-language explanation.

That said, visualizations can be built with tools designed for that purpose (Tableau, Power BI, Looker), with more generic presentation tools (PowerPoint, Google Slides) or with Python / R / JavaScript (through the appropriate libraries). In that respect, the advisable approach is to adopt one in particular and go deep into its use.

Further reading

Visualizing amounts, proportions and distributions

Basics of storytelling and effective presentations