How to start a visualization task

We’ve all been in this situation. The situation in which you have a load of data available, and you need to start diving into it. For example:

  • You’re a scientific researcher who just finished an experiment. You need to analyze the data and compare to other studies to verify your hypothesis.
  • You’re an online marketer who just ran a great marketing campaign. Your boss asks you for the key learnings to optimize future campaigns.
  • You’re the CEO of a small company. You need to monitor your pipeline on a daily basis to stay up-to-date.

Visualizing your data will help you to detect the insights you need, and present them in a clear way. Now, how do you get started on that visualization? This step-by-step guide will help you bring the task to a good end!

3 types of data visualization

There are different ways to approach data visualization. Your approach will depend on the type of data, and the goal you want to achieve.

In general, we distinguish 3 types of data visualization. Each type is fit for a different goal, and comes with its own difficulties. Before starting your visualization task, assess in which of the following stages you are.

3-types-data-viz


1. Discovery

This is the most complex task. When in the discovery phase, it means you basically received a data dump. You’ll have to create order in the large pile of data.

Some examples:

  • Investigative journalism
  • Research & academics
  • Anomaly detection

What are the difficulties?

iconmonstr-database-10-240.pngData. Lots of data. We’re not talking megabytes. Rather think on a GB, TB or even PB scale. This implies that you can create thousands of hypotheses. Due to the scale, data quality is often very poor. In addition, this often brings along high-dimensional data, meaning your data samples have a very large number of dimensions, which can make it painful to work with.

How to get started?

Try to get a grasp on the data you’re working with. Start by understanding how it was collected or produced. Has the collection method introduced biases that may influence your analysis?

For example, think of a call center that only calls on working days, and only to land-lines. Chances are that a lot of your calls will be answered by retirees.

Once you know more about the collection, start visualizing each dimension & measure separately. Get a feel for what your data looks like: how are they distributed? Are there many missing values?

Finally, unleash more powerful visualization platforms to look for correlations, and mine for insights!


2. Persuasion

Data visualization is an extremely powerful form of storytelling. When you need to present an idea, your message will be much stronger if you have the necessary data as proof. With the right visual representation, you’ll convince your audience in minutes. No more needless discussions on conflicting interpretations.

Some examples:

  • Regular journalism
  • Policy & politics

What are the difficulties?

presentation-1559937_1280You need to make sure that your story will convince a wide audience. At the same time, you have to be aware that data visualization is never objective. In some way, you’ll always emphasize certain aspects and overshadow others. Therefore, it can be tough to stay honest and persuasive at the same time.

How to get started?

First of all, always keep your focus on the insight that you want to give. Make sure that your visualizations are a reinforcement of the message you’re trying to convey. Secondly, use a graph type that is fit for your message. Check out this simple flowchart on which graph to use. More tips on how to persuade your audience? Download our whitepaper full of best practices.


3. Process

Data visualization can help companies all over the world to stay on top of their business. Follow the daily status or evolution of your business with real-time dashboards. Those dashboards serve as a daily guidance for decision-making.

Some examples:

  • Manufacturing industry
  • Marketing & Advertising industry

Basically, any company worldwide!

What are the difficulties?

iconmonstr-monitoring-1-240.pngWhen you’re monitoring certain KPI’s on a daily basis, you need to ensure that you aren’t missing any big shifts. It could be that other factors – which you aren’t monitoring yet – are influencing your results, without being aware of this.

How to get started?

When setting up your dashboards, make sure to define the KPIs that really matter to your business. Try to keep it simple at all times. If variables are connected, they need to be seen together. If they are not, keep them separated to avoid confusion.

From time to time, take a step back from your current dashboard. Think of other variables that may be influencing your numbers, and start taking them into account.


Identify in which stage you are

When starting a visualization task, first think of the goal you’re trying to achieve. This will help you look at your data in the right way. It will ease the visualization process. With these tips to get started, and a warning for the possible traps, you’re all set to start working more efficiently with your data!


Start visualizing your data!

Add a Comment

Your email address will not be published. Required fields are marked *