Data visualisation is one of the most popular and useful method of data analysis. It gives a visual representation of the variables and their relations unearthing more information that could have easily been missed or difficult to derive form just looking at the data. Variables can be numerical or categorical. The choice of visualisation techniques depend on the type of variable(s) in question and the goal of the visualisation. There are a lot fancy colors that can be used in visualisations, some are give the visual more clarity but others obscure the results. Clarity above aesthetics is the rule of visualisation. There are three main type of analysis:
-
Univariate analysis - this is the analysis of a single variable. Some of the most commonly used visualisations for these are;
- Histograms
- Countplot
-
Bivariate analysis - this entails the investigation of two variables on how they interact. Some of the most commonly used visualisations for these are;
- Scatter plot
- Boxplot
-
Multivariate analysis - this is the analysis of more than two variables and their interactions. Some of the most commonly used visualisations for these are;
- Heatmap
- Pairwise plot
They are many more ways of visualising data, some of which cut across the 3 methods eg barplots can be used in any of the analysis with categorical variables. Remember the choice of visualisation is heavily governed by the type of variable and intended goal of the analysis. Some variable could be numerical representations of geographical locations and as such using a countplot would produce little information eg frequency of a location which is barely any information. in this kind of scenario, a map would be the most ideal visualisation, coupled with other variables would be very informative. The most important thing to remembe during visualisation is to try and be as clear as possible, let the visualisation tell pass the information as concisely and effortlessly as possible.
Happy visualising!
Top comments (0)