Data Visualization Best Practices : Part 1 – Choosing the right colors for bar charts

Bar charts are the most commonly used charts in Business Intelligence reports and Dashboards. They are useful in summarizing the categorical variables or qualitative variables where the length of the bar is directly proportional to value they represent. This post is one of many posts related to best practices in data visualization. In this post we will learn which color pallete should be used in bar charts.

Let us assume a company is manufacturing 4 products namely A,B,C,D and your manager want to see the sales of these product during 2012 using a bar chart.Which bar chart would you choose?

bar

In the first bar chart all the four bars are represented by four different colors. Most of the people will be tempted to use this bar chart as it looks attractive.Actually the colors in the bar do not convey any meaning and tends to confuse the user as user searches for meaning for each color resulting in wastage of time.

Color scheme followed in the third bar should be avoided as this color scheme is best suited for plotting the heatmap. In this color scheme order plays an very important role. Color with lower shades represents a variable with less value and color with high intensity represents the variable with high value. So this color scheme should be avoided in bar charts.

Based on the best practice second bar chart is recommended as all bar charts are represented by same color, It makes easy for user to compare values across all bars.

5 thoughts on “Data Visualization Best Practices : Part 1 – Choosing the right colors for bar charts

  • Aside from the choice of colour logic, is there an argument to say that the data would be better presented with a horizontal bar chart, with same colours as advised above, as in reality the Product instances of A,B,C,& D would in fact have much longer identifiers / descriptors and these tend to be difficult to neatly reflect on a vertical bar chart. Also, with a ranked horizontal bar chart, i.e. from biggest seller to smallest seller, the horizontal category instance lables would form an ‘easy on the eye’ ranked list from which the User could quickly understand the relative sales of each Product. What happens when the count of Product instances goes into the higher numbers of 10, 20, etc.

    Just a thought…

    Good post
    Paul C

Leave a Reply

Your email address will not be published.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