My Journey Through the Best Data Visualization Practices
In today’s data-driven business landscape, organizations accumulate large amounts of data every day that can provide valuable insights. As a data analyst, it is my responsibility to interpret this data through statistical analysis and present it to stakeholders in the form of data visuals and dashboards. Effective data visualization is crucial as it simplifies complex information into easily understandable charts, enabling business owners to make informed decisions quickly.
However, to effectively communicate these findings, I find it essential to adhere to certain best practices which I have accumulated through the trials and errors in my data analytics and visualization journey.
In this article, I will be doing a rundown of an analysis I carried out earlier this year on sales data. This was my first dashboard on Power BI from which I learned a lot, and from these methodologies, we will look into the best practices in detail. You can download the dataset here and practice later on after you’re done with reading.
About the data
I’ve used training data stimulated from real business data after taking out some vital information. To read more about the analytical firm that created the data, and to check out other datasets for your portfolio projects, click here.
Foresight Pharmaceuticals is one of the leading pharmaceutical manufacturing companies with a global presence. Their markets are divided into different regions across the world. They have an agreement with their distributors to share their sales data with them. This is to enable them to gain insights up to the retail level.
Now let’s get down to these best dataviz practices using the case study.
1. Set Clear Visualization Goals
Before diving into the creation of dashboards, I tend to establish the purpose behind my visualizations by defining the key messages I want my audiences to take away from the visuals. This thus will guide the design and the content of my dashboard, ensuring a cohesive and impactful presentation.
Given the information on Foresight Pharmaceuticals’ company as my case study, first, I tried to understand what my data consisted of (understanding the domain knowledge) and then generated problem statements using the following questions :
- How does the sales performance vary across different regions or territories?
- What is the overall sales generated by the sales team during a specific period?
- Who were the top distributors of these products and services across the years?
- What are the top-selling products or services and their respective sales figures in the German and Poland markets?
These are only a few questions that will define my goals for this analysis.
2. Understand Your Audience
By understanding my audience, I tailor my visualization to effectively communicate the insights that matter most to each group.
- For the executives, I put my focus on presenting high-level KPIs, such as overall return on investment (ROI) and revenue impact.
- For the marketing managers, I can dive deeper into metrics like conversion rates, customer acquisition costs, and campaign performance across different channels.
Based on my sales data from the case study, I decided to show the following metrics:
- Total sales.
- The yearly trend of these sales across the different countries.
- The products that generate the most sales.
By aligning my data visualization with the specific needs and interests of my audience, I enhance their understanding and engagement with the data. This is done to ensure that my visualizations are relevant, impactful, and can effectively communicate the story behind the data to drive informed decision-making.
3. Gain a Deep Understanding Of Your Data
A thorough understanding of your data and its requirements allows for the creation of accurate and meaningful representations of information. It helps identify patterns, trends, and relationships that can uncover valuable insights.
In the data preprocessing stage, I eliminate and clean my data from any nuances or outliers that could skew the insights. Understanding my data ensures the credibility and reliability of visualizations, leading to more impactful and actionable outcomes.
4. Choose the Right Visuals
Different types of charts serve distinct purposes, and it’s crucial to familiarize yourself with the applications of the various visualization options to convey the most accurate and appropriate message to your audience. Selecting the right visualizations will enhance comprehension and aid effective decision-making of your audience.
The charts below are graphical representations of understanding your data visuals and when to use them. I wanted to visualize the yearly trend of sales of the two countries (Germany and Poland).
In this case, a pie chart would be wrong because it primarily represents proportions or percentages of a whole. It lacks the ability to effectively display changes over time or compare values between different categories. Line or area charts are used to visualize changes over time, and you can clearly get your insights at a glance. Just like in the chart below, you can notice the peak in 2019 and can easily tell the difference for the other years.
5. Simplify and Reduce Clutter
I strive for simplicity and avoid overwhelming my visualizations with unnecessary elements. To do this, I carefully consider my placement of axes, visuals, and data labels, ensuring they contribute to the clarity of my message. Eliminating clutter allows my audience to focus on the essential insights, improving overall comprehension.
For an illustrative example, I will refer to another project of mine, dedicated to Sports.
Here’s a question for you. What can you deduce at first glance from the chart below? Imagine having a dozen of this type of chart on your dashboard.
During my analysis of the Women’s FIFA World Cup from a project I worked on earlier this year, I had a goal to illustrate the annual trend of team performances in terms of goals scored. The provided chart above includes various components such as axes, data labels, markers, and series labels, showcasing the yearly trends for all the teams. However, it is important to note that in this scenario there may be a tendency to prioritize the inclusion of multiple elements in the visualization while overlooking its overall effectiveness.
By examining the chart above, could you identify any noticeable differences or changes?
My specific inquiry focused on understanding the trends of the top three teams with the highest goal count, as I aimed to compare them. To accomplish this, I simplified the visualization by removing unnecessary chart elements and filtering out countries that were not relevant to my analysis. Additionally, there could also be an option to visualize the trend of a single country based on your specific problem statement.
6. Choose Colors Strategically
But let’s return to our sales data.
The effective use of colors can significantly enhance data visualization. Select a consistent color template and format that aligns with the context and purpose of your message. Avoid excessive use of colors, as it may lead to confusion. I often opt for a minimalistic approach, ensuring the colors I choose accurately convey the intended meaning. It would be a hassle to have different charts on your dashboard with several color themes, which most likely will lead to color riots which can cause confusion and an inability to focus on the key insights for the audiences.
For business dashboards, it is better to use a minimalist color palette to avoid distracting users with a variety of colors. Here is an example of a restrained, but still not optimal, color palette.
For different data categories, sometimes it’s better to use the same color in charts like bar charts. Another successful technique we will explore in the next section.
7. Highlight Key Insights
One technique I always use while creating my dashboards is making my important insights stand out from the rest of the visual elements. I do this by utilizing conditional formatting processes to emphasize crucial information. This helps your audience quickly identify and comprehend the key takeaways, facilitating informed decision-making. From the chart below, it’s very easy to identify that Britanny Bold is the manager whose team generated the most sales at 30.84%.
As a result of my trials and errors, my first dashboard looked like this. Of course, now I look at it with love and tenderness, just like at my first creation. Although I understand that it can be improved in various aspects.
Since then, I have created several more dashboards on different topics, and during their development, I tried to take into account the mistakes from the past and apply the best data visualization practices. I hope that in the future, my projects will become even better!
8. Think Like a Designer
The last recommendation that I couldn’t fully implement in my early projects, but over time, I realized the importance of developing aesthetics as well.
Now I usually approach my data visualization with a design mindset. It’s important to be flexible and adaptable to different tools and platforms, ensuring your visualizations are aesthetically pleasing. Impressions matter, and a visually appealing presentation enhances understanding and retention of information.
One of the things I do before starting my dashboard is making a wireframe and sketch of what I want my dashboard to look like: the general concept, the placements of different charts, and chances of modifications. At times it may be hard to come up with a concept. I’ll share what I did in my recent Power BI dashboard.
- Since it was a project on the analysis of Tweets about a song album, I wanted my dashboard to contain similar colors to the album cover. Using the color picker I extracted colors from the album cover and applied them to my visuals.
- I searched different samples of dashboards to get inspiration for creating mine. Finally, I got dark samples on dribble and it guided me on my own creation.
This was my final dashboard after several days of editing and modifications. To read more on the process methodology I applied for the project click here.
Another approach I suggest is having a collection of beautiful dashboards that inspire you and go through them from time to time. Through this, you can understand the design perspective of people, and learn a few from their visuals. Most importantly, we should be free to recreate what’s on our minds in the best possible way. Using this site you can get color palettes, themes, and it also gives you the opportunity of having complimentary colors for your dashboard.
9. Don’t Forget to Tell Stories about Your Data
The art of data storytelling is significant in data visualization. It can help engage and captivate the audience, enabling them to connect with the data on a deeper level. The following below are chart types and their applications, and with them, we can build compelling visuals to convey our messages to the audience:
- Bar Charts: Ideal for comparing categorical data or showing frequency distributions.
- Line Charts: Effective for illustrating trends and changes over time.
- Pie Charts: Useful for displaying proportions or percentages in a whole.
- Scatter Plots: Suitable for showcasing relationships between two variables. But remember, for a broad audience, the perception of this chart may be challenging.
- Heatmaps: Great for presenting large amounts of data and identifying patterns.
- Treemaps: Helpful in displaying hierarchical data structures.
- Funnel Charts: Effective for visualizing the progressive stages of a process.
- Area Charts: Ideal for showing the distribution of categories as part of a whole over time, where cumulative is unimportant.
In addition to choosing the right charts, use color highlighting and annotations to tell your story in the best possible way.
To explore different dashboard styles, diagrams and to understand how people use them, you can check out Tableau’s viz of the day. In this gallery, there are always many interesting examples of storytelling and dashboards.
Overall, by adopting these aforementioned recommendations, you can achieve a huge breakthrough in creating excellent visualizations and dashboards.
Now you have an idea of the common pitfalls to avoid, go ahead and try yours!
Here’s a quick summary about me:
My name is Nnodi Precious. And I’m an engineer-in-view with a strong passion for creating solutions for businesses through insights from data visualization and dashboards. In my free time, I love to read about new trends and technologies in business data analytics.
And I would love to connect on LinkedIn.
