When I started my career I was curious to understand the 3 W’s — What, Why and Where — of dataviz. Now, after enabling multiple businesses in India to consume data through visualisation, I have found a few fairly convincing answers that may help someone getting started in their data visualisation career. Let’s dive in:
‘What’ is Data Visualisation?
Okay, that was too much marketing content, mostly about the importance of a specific dataviz tool. Great content for a sales executive. But none of these would help a newbie getting into dataviz to understand the basic question: — “what” is dataviz.
So, from the eyes of a dataviz practitioner, “what” exactly is Data visualisation? Data visualisation is a process which consumes data as input and transforms it into business insight (or) stories
Charts by themselves do not provide insights or stories directly.
We achieve it through further steps, such as annotating the chart to highlight key aspects. You can learn more about annotations by listening to Episode 07: How to Annotate Like a Boss! Featured Viz by Susie Lu — Alli Torban’s Data Viz Today.
Organisations have two things in abundance: data and unanswered questions.
Organisations have two things in abundance: data and unanswered questions. Decisions have to be made based on the answers to those questions. Dataviz enables them to convert data into insights which help organisations to make data-driven rather than gut-based decisions.
‘Where’ is Dataviz used?
To measure performance, most businesses track a series of Key Performance Indicators (KPI) through reports or Management Information Systems (MIS). As data storage becomes less expensive, organisations collect (or) buy as much data as possible to come up with KPIs based on the collected data
It is not easy for decision-makers to consume a large amount of data through traditional reporting systems and derive actionable insights. This is where dataviz comes in handy. It makes the decision-makers life simple by converting a large amount of data into business insights through visuals form and design
There’s one H too: ‘How’ to derive insights as business stories from dataviz?
There are any number of dataviz options, styles, and approaches that are available to the practitioner. It is critical to always know 2 things about each viz:
- The input data (dimensions & metrics)
- The potential business insights (or) stories which could be derived from it
Case Study: When to use a line chart & when to use a scatter plot?
Line charts and scatter plots are pretty similar. Let’s understand what could be the “input” data and what could be the “output” business insights and stories. We plot a metric (mostly the dependent variable) against the dimensions in the X and Y axes. The difference between a line chart and a scatter plot is just the way in which the “line” is created.
Line chart: Individual data points are connected by a line. A line chart is used to track changes over a period of time, to identify if there is a pattern being formed over a period of time
Scatterplot: Does not connect individual data points, but shows a “trend” of the data points. Scatterplots are used to figure out the relationship between 2 variables through a slope (regression line) — for every unit change along the X-axis, what is the respective change on the Y-axis. By doing this we can also find out the presence of outliers and why they are important. For more reading, check out: The power of outliers (and why researchers should ALWAYS check for them).
What could be the input?
- Number of dimensions: 2 (one in x-axis and one in y-axis)
- Number of metrics: 1 (minimum)
- X-axis: Independent variable (Eg: time)
- Y-axis: Dependent variable (Eg: inflation, population, TV rating point, etc)
In order to derive insights as a story from a scatterplot, look at the slope. Based on the example shown here on the left, from the slope, we infer that for every additional hour spent studying, the student’s score goes up by 15 %
In order to derive insights from a line chart you can use an average line. Based on the example shown on the left, we infer that TV ratings have been declining over the past 2 weeks and stay below the 13-week average. Annotations could be used to talk about critical points above and below the average line as shown in the example.
Dataviz is a framework to consume, communicate & convert data into memorable stories or insights. These insights enable businesses to answer questions and make better decisions.
For someone getting started, I would suggest to clearly understand the unanswered questions and relevant data first, then consider a viz which would provide answers to the questions by consuming the relevant data. Finally, it is critical to learn how to derive insights as stories from the chosen viz that answer questions based on which decisions would be made
Krishna is a lead Data Consultant at Gramener. He has been enabling businesses to consume data through design and consulting which foster key stakeholders to answer real-world questions. He has been providing dataviz solutions across TV media and marketing domains.