Art of Visualisation
Visualisation of data is a art by itself. Data which is mostly present as rows and columns in a table/csv file have to be best visualised to better understand the phenomenon underlying the data.
Visualisation brings data to life
Sometimes it is easy to spot trends, seasonality, noise and other characteristics of the data/signal by just visually looking at the data. Thus visualisation plays a vital role for doing data analysis and more predominantly during reporting.
I always visualise/plot the data/signal before performing any analysis to feel the data. Some of the basic questions I ask with the data in hand first is that
- Is this data sampled at regular intervals of time (uniform data)?
- Is this data structured?
- What is the underlying process/phenomenon that generates this data?
- How this data will help me answer the problem that I want to solve?
- Does the data contain any noise? Is so, what kind of data pre-processing steps that I have to perform to clean the data?
As a Data Analyst, we have to understand the business problem first. Understanding of business problem leads to choosing right data to use for solving the business/organisational problem. Most of the time, the business/organisation collects plethora of data most of the which are not suitable for the solving a particular problem.
Data can be discrete or continuous. For example, if we are collecting data about cyclist crashes, then the data will be discrete. If we are collecting data about temperature/air quality the data will be continuous.
Graphs are mainly used for visualising continuous data whereas bar charts are mainy used for discrete data. The main important question we have to ask is that whether visualisation can reveal much information about the data.
“Can we intuitively understand the information that the data wants to convey” should be the main focus of data visualisation.
There are plenty of analytics tools that can automatically visualise the data. However as a data analyst you have to be careful with them. Being carried away with some automatic recommendations may lead to unsatisfactory outcomes of the business problem that you want to solve.
In the nutshell, data visualisation is very indispensable for solving a business problem and ability to articulate the outcomes in a beautiful way