What are the most common scatterplot mistakes?

1 Wrong scale

One of the first things to check when making a scatterplot is whether the scale of the axes is appropriate for the data. A wrong scale can distort the shape and distribution of the data, hide important details, or exaggerate minor differences. For example, if you use a linear scale for data that has a large range of values, you might end up with a plot that is squeezed on one end and stretched on the other. A better option would be to use a logarithmic scale that can accommodate the variation in the data. Similarly, if you use a scale that is too narrow or too wide, you might miss some outliers or create unnecessary white space. A good rule of thumb is to use a scale that covers the minimum and maximum values of the data, with some margin for readability.

2 Wrong aspect ratio

Another aspect that can affect the appearance and interpretation of a scatterplot is the aspect ratio, which is the ratio of the width to the height of the plot. A wrong aspect ratio can change the perceived slope and correlation of the data, making it look stronger or weaker than it really is. For example, if you use a tall and narrow plot for data that has a positive correlation, you might overestimate the strength of the relationship. A better option would be to use a square or a wide and short plot that can show the true correlation more accurately. Similarly, if you use a square or a wide and short plot for data that has a negative correlation, you might underestimate the strength of the relationship. A good rule of thumb is to use an aspect ratio that matches the range and variation of the data, or to use a standard aspect ratio of 1:1 or 4:3.

3 Wrong color

Another element that can enhance or detract from a scatterplot is the color of the points. A wrong color can make the plot hard to read, confusing, or misleading. For example, if you use a color that is too light or too dark, you might lose contrast and visibility of the points. A better option would be to use a color that stands out from the background and the axes. Similarly, if you use a color that is too similar or too different from other elements in the plot, you might create ambiguity or distraction. A good rule of thumb is to use a color that is consistent and harmonious with the theme and purpose of the plot, and to use different colors or shapes to distinguish different groups or categories of data.

4 Wrong size

Another factor that can influence the readability and impact of a scatterplot is the size of the points. A wrong size can make the plot cluttered, sparse, or misleading. For example, if you use a size that is too large or too small, you might obscure or hide the underlying distribution and variation of the data. A better option would be to use a size that is proportional and appropriate for the scale and density of the data. Similarly, if you use a size that is not related to any meaningful variable or attribute of the data, you might create confusion or false impressions. A good rule of thumb is to use a size that is meaningful and informative, and to use different sizes to represent different values or levels of a variable.

5 Wrong label

Another component that can improve or ruin a scatterplot is the label of the axes, the points, or the plot itself. A wrong label can make the plot vague, inaccurate, or irrelevant. For example, if you use a label that is too long or too short, you might create clutter or ambiguity. A better option would be to use a label that is concise and descriptive. Similarly, if you use a label that is not aligned or consistent with the data, you might create errors or inconsistencies. A good rule of thumb is to use a label that is clear and accurate, and to use different labels to identify and explain different aspects or features of the data.

6 Wrong message

The last and most important thing to consider when making a scatterplot is the message that you want to convey with the data. A wrong message can make the plot irrelevant, misleading, or biased. For example, if you use a message that is not supported or justified by the data, you might create false or spurious conclusions. A better option would be to use a message that is based on sound and valid analysis and inference. Similarly, if you use a message that is not relevant or interesting for the audience, you might create boredom or indifference. A good rule of thumb is to use a message that is meaningful and engaging, and to use different techniques to highlight and communicate the key insights and findings of the data.

7 Here’s what else to consider

This is a space to share examples, stories, or insights that don’t fit into any of the previous sections. What else would you like to add?