Good data visualizations are a powerful way to communicate insights from data. They can make it easier to derive information and meaning from data; reducing the cognitive burden of reviewing, processing, and understanding the insights your data has to offer.

 Before we start thinking about how to build a visualization, it's important to get to know the data we are working with,. This step is sometimes referred to as exploratory data analysis. There are a variety of tools and techniques that we can use to do this, but one of our favorite methods for exploring data is by creating a profile of the data. This can help identify patterns and trends that may not be apparent when looking at raw data. At Edge Analytix, we love the Pandas Profiling library to easily create in-depth profiles data sets. The profile page includes a breakdown of your data types, a summary of important descriptive statistics such as mean, standard deviation, and skewness, as well as histograms of each field to help you visualize the distribution of the data. We'll cover profiling of data sets more in-depth in an upcoming blog post!

 Another important step toward creating data visualizations is to ask the right questions. We usually ask ourselves the following questions before even beginning to draft the first visualization:

1. What is the context for these data?

2. Who is the audience for the visualization?

3. What message do I want to get across with this visualization?

4. How do I want my audience to feel when they finish the visualization?

5. Where will the visualization live?

Once we have a good idea about the context and purpose of our visualization, we can then start to think about the types of visualizations we can create. There are many different types of visualizations that can be created depending on the data and message we are trying to convey. These can include line graphs, bar charts, scatter plots, pie charts, etc. I'll admit that I'm biased against visualizations named after foods (pies, donuts, and spaghetti), but even these visuals can be useful when telling the right story and in the right context. When creating a visualization, we always try to consider the most important elements we (or our customers) want to highlight and what our audience needs to know about the data being represented. Once we've settled on a visualization that meets our needs, we start sketching it out to make sure that the message is clear and the design looks good. Once the sketch is finalized, we can start creating the visualization in a tool like Tableau or Python.. During this process and while we work to finalize our visualizations, we always try to keep in mind six fundamental characteristics of good data visualizations:

1. Simplicity - Simplicity increases comprehension. Complex visualizations make it difficult for viewers to understand and retain key insights. The first rule we try to follow when developing our visualizations is to make them as simple as possible while still retaining the key details that are important to the data.

2. Visual Appeal - Data visualization should be pleasing to the eyes. Viewers are less likely to understand the message if it is difficult or unpleasant to look at. Simple shapes with clear contrast often work the best and they are most likely to catch they eye of our viewers when we need to highlight certain elements. Our approach to visual design is to aim for simplicity and elegance over flashiness.

3. Credibility - Many visualizations rely on graphs and charts that readers have encountered before. By using visuals that we are familiar with, it is easier for viewers to process the information and remember the key takeaways from the visualization. Our approach has been to stick with the most commonly used types of data visualizations and try to leverage existing conventions as much as possible.

 4. Relevance - A good data visualization should highlight key points that are relevant to our audience's goals. Every visualization should serve a purpose, and supporting that goal with well-designed visuals is key. As data analysts, it's easy to fall into the trap of highlighting what we might find interesting about the data, but that may not always be consistent with the insights that are important to adding value to the business. Making sure that our visualizations align with the goals of our customers is essential in creating effective data visualizations.

5. Consistency - The visual components of a visualization should match and be consistent with each other. This repetition and consistency helps viewers find meaning in the messages you are trying to convey. When a viewer visits a new visualization, if the layout, style and branding feel consistent, it can effectively reduce the learning curve for understanding the meaning of the data.

6. Flexibility - The best data visualizations should evolve as needs change. We aim to build things in a way that components can be reused and modified based on feedback from customers. By remaining flexible, we ensure that our designs will support our customer goals. Keeping flexibility in mind from the beginning of a project can often reduce the costs associated with developing additional visualizations and reduce both time and cost to produce additional products for our customers.

These six principles have been critical in helping us create successful data visualizations for a variety of clients across a variety of business sectors. Whether you're an analyst, a designer, a customer, or even just a casual reader, we hope you will find this information interesting and useful in designing effective data visualizations. Thanks for reading!