Data visualisation is the process of representing data graphically. A data set represented visually speaks more than just numbers. The human brain is wired to identify itself with visuals rather than numbers. Thus, a pic of a pile of mangoes tempts us more than the raw data point of “there are 100 mangoes in the basket”.
Visualisations include charts, graphs, tables, maps and many more. These elements together in an organised way help users to identify patterns and hidden relationships. The objective is to create a narrative story that benefits the decision-makers
Personally, I have seen that clients are not as much moved by charts and graphs as they are moved by insights. Typically, when we show them the data, the response is lukewarm. But if we show them insights, they give us business.
Dos of Tableau Visualisation Don’ts of Tableau Visualisation
Understand the data Avoid complex designs
Know the end-user Designing without understanding the data, scenario and user
Keep simple designs pattern Inappropriate chart selection
Validate report Crowded KPIs in reports
Maintenance of Report Throwing away random chart colours
I have seen that my enterprise clients may have a knack for business, but not all of them all the time are keen on playing with numbers. Frankly, they may not have time for that. So, we avoid complex designs wherever possible.
The charts or data representation should have to be easily comprehensible. If not, their purpose is defeated, before they can be interpreted. The data visualisations need to be simple, ultimately to maximise their value to the end-user.
Also, the schema of the business visualisation differs from domain to domain and specifically from user to user. And sometimes, what suits one department may not suit the other department.
From my experience, I can state that standard colour palettes are the best ones to be followed. Ultimately, they have to be sensible and pleasing to look at. The fundamental aspect is to keep the noise to a minimum. One has to be careful in colour selection for the dashboards. And when creating reports, for instance, a report on a department’s performance vis-à-vis other departments, one has to be consistent with the use of colours. For example, if you denote the HR department in a particular colour, it is advisable not to switch or change the colour if it is referred to elsewhere.
Tableau enables big data analytics by providing connectivity from Big Data platforms such as Hadoop, Spark and No SQL databases to cloud data warehouses. By extracting the data to Tableau, we can build charts to identify patterns and trends.
Tableau, as far as I look at it, is good in data-visualisation gymnastics! It is so flexible that it can have connections to multiple Big Data platforms simultaneously and then solicit data from those platforms for representational or visualisation purposes. Thereafter it is as easy as 123, depending on the complexity of the project at hand.
The following methods are used to improve performance:
Reducing the no. of sheets
Using extract
Avoiding unnecessary fields
Reducing the number of filters
Reducing the number of string calculations
Imagine there are 5 Sheets, and they exhibit 5 different KPIs. The client is happy when they could be merged into one Sheet with Insights in, rather than five sheets separately shown.
The clients want maximum KPIs with a minimal number of data exhibits, to facilitate maximum performance inputs. They are keen to preserve efficiency and effectiveness, no matter what. It helps them stay on top of the situation, however hard or easy it turns out to be.