Nowadays, data visualization is the main topic in the field of data science. However, even if you try to make a nice presentation or a report, the message is not always delivered well. It is easy to get lost in many numbers, charts, and labels. In this article series, we will share some tips about data visualization and give some advice on creating graphs. Let’s start with bar charts.
Bar Charts
Bar charts are one of the most common types of data visualization. They are mainly used to show changes over time or to compare different categories and part-whole connections. Here you can see four types of bar charts and the usage fields.
Vertical: Mostly used with timelines
Horizontal: Mostly used with long category names
Stacked Bar Chart: Mainly used for part-whole connection
100% stacked diagram: Mainly used to test the dispersion
Bar Chart Guide
What should you avoid when using bar charts? How can you make more your data understandable? Here are five tips.
**Tip 1 ** To make sure the labels are readable, use horizontal labels.
Tip 2 The optimal distance between the bars should be half the width of the bars.
Tip 3 Start the scale at zero.
Tip 4 Use only one color. Also, it is better to avoid stacked bar graphs because they are hard to understand at first glance. However, if you use them, use different colors.
Tip 5 Alphabetical order makes the data easy to find.
Summary
Bar Charts
- Use horizontal labels for readability
- Keep the optimal distance between bars (half the width of the bars)
- Start the scale at zero
- Use only one color (unless you are using a stacked bar graph)
- Alphabetical order makes the data easy to find
In addition to all of the tips above, there are other rules we need to keep if we want to create correct graphs. However, following these tips can improve the quality of your bar charts and presentations.
Bonus Tip : Always try to find the best type of bar chart or any graph that fits the data and situation perfectly.
Source: http://www.slideshare.net/Visage/data-visualization-101-how-to-design-chartsandgraphs
