• Python Snacks
  • Posts
  • Visual biases: Understanding Biases Within Data Visualization

Visual biases: Understanding Biases Within Data Visualization

See how you can improve your plots by understanding visual and cognitive biases.

Before getting into this week’s snack…

In the previous newsletter, I discussed the cognitive biases of data visualization. This topic covered 3 kinds of biases: confirmation, anchoring, and framing bias.

Cognitive biases are these “mental shortcuts” that can lead people to think or act in certain ways that aren’t always logical or rational. They affect how we perceive, interpret, and remember information.

From a business standpoint, this can lead to flawed decisions. So, by being aware of both cognitive and visual biases, we can make better visualizations and more informed decisions.

Visual biases in data visualizations

Suppose that you are reporting the quarterly sales performance of three different software products over a year to management. The data collected is as follows:

Quarter

Product A

Product B

Product C

Q1

200

300

250

Q2

220

320

270

Q3

210

310

260

Q4

230

330

280

Without any biases, we’d anticipate the graph to look something along these lines:

No biases are introduced in this graph

Through this plot, we are trying to communicate to management:

  1. Overall sales trends

  2. Comparative quarterly performance over 3 products

  3. Identifying consistent performance

Let’s see how 3 kinds of biases can affect what we’re trying to communicate to management.

Scale Manipulation

Scale manipulation occurs when we adjust the x or y axis to visually manipulate the data. In this scenario, we are “zeroing” the y axis, which isn’t necessary for this situation.

This can mislead viewers by making small differences appear more significant or larger differences appear trivial.

Confirmation Bias - Top graph showing bias, bottom graph without bias.

Management interpretation with bias: Management may see consistent sales across all 3 products despite some products potentially doing well, as the graph is flat in comparison to the one where there is no “zeroing” of the y-axis. This could lead to underestimating performance of a product.

Management interpretation without bias: Management would understand the actual scale of differences between quarters, recognizing that the changes in sales are relatively minor and part of normal sales variation.

What this graph fails to communicate: This graph fails to provide a truthful representation of the data, potentially leading to overreactions or misinformed decisions based on exaggerated visual cues.

Color Bias

Color bias occurs when the use of color in a graph highlights certain data points over others, which can draw undue attention to specific aspects of the data and mislead the viewer.

Graph introducing color bias

While yes, you have a pretty colorful graph, those colorful points serve no purpose!

Management interpretation with bias: Management might focus more on the products with brighter or more distinct colors, assuming these products are more important or have more significant sales patterns than others.

Management interpretation without bias: Management would view all products equally, understanding the data without being influenced by the color scheme and making more balanced decisions based on the actual numbers.

What this graph fails to communicate: The graph with color bias fails to present an impartial view of the data, potentially leading to skewed interpretations and decisions that are influenced by the visual emphasis rather than the data itself.

Overplotting Bias

Overplotting happens when too many data points are plotted on a small graph, causing them to overlap and obscure important information.

In this scenario, we are trying to plot daily sales on top of trying to show the quarterly sales.

Framing bias - User Base and Engagement

Management interpretation with bias: Management might find it challenging to interpret the data due to the cluttered appearance of the graph, potentially missing important trends or fluctuations in sales performance.

Management interpretation without bias: Management would clearly see the overall trend of sales performance over the quarters, understanding the fluctuations and making informed decisions based on the clear presentation of data.

What this graph fails to communicate: This graph fails to provide a clear view of the data, obscuring important details and making it difficult to identify trends or significant changes in sales performance.

Take-Home Points

So between cognitive and visual biases, here are 3 tips I want to leave you when you go to make your next graph:

  1. Accurate Data Presentation is Crucial - always ensure that your data visualizations are free from biases such as scale manipulation, overplotting, and color biases. Accurate and clear data presentation helps in making informed decisions and avoids misinterpretations that can lead to strategic missteps.

  2. Understand the Impact of Visualization Choices - The way data is visualized can significantly influence the viewer's perception and interpretation. Choosing appropriate scales, avoiding unnecessary clustering or overplotting, and using consistent color schemes can prevent misleading interpretations and ensure that the true story of the data is communicated.

  3. Regularly Review and Validate Graphs - Before presenting data to management or stakeholders, review and validate your graphs to ensure they accurately represent the data. This includes checking for any potential visual biases and confirming that the graphs effectively highlight the key insights and trends necessary for decision-making.

📧 Join the Python Snacks Newsletter! 🐍

Want even more Python-related content that’s useful? Here’s 3 reasons why you should subscribe the Python Snacks newsletter:

  1. Get Ahead in Python with bite-sized Python tips and tricks delivered straight to your inbox, like the one above.

  2. Exclusive Subscriber Perks: Receive a curated selection of up to 6 high-impact Python resources, tips, and exclusive insights with each email.

  3. Get Smarter with Python in under 5 minutes. Your next Python breakthrough could just an email away.

You can unsubscribe at any time.

Interested in starting a newsletter or a blog?

Do you have a wealth of knowledge and insights to share with the world? Starting your own newsletter or blog is an excellent way to establish yourself as an authority in your field, connect with a like-minded community, and open up new opportunities.

If TikTok, Twitter, Facebook, or other social media platforms were to get banned, you’d lose all your followers. This is why you should start a newsletter: you own your audience.

This article may contain affiliate links. Affiliate links come at no cost to you and support the costs of this blog. Should you purchase a product/service from an affiliate link, it will come at no additional cost to you.

Reply

or to participate.