Visualize Your Data: Making Sense of Data Chapter 4

Chapter 4 covers some basic charts that will be used to visualize data. I think most of my readers would know those charts very well. But, I would still like to recap some important points. Still, I learned some new things as well, such as Pareto Charts, and recalled some old memory in high school where I plotted “stem-and-leaf plot”.

Let me first introduce Pareto Charts which is new to me. And then I will provide a summary on different charts.

At a first glance, it is like a bar chart. But, the categories are arranged in descending order from left to right. Pareto is actually the one who discovered the 80/20 rules. Therefore, in Pareto Chart, the cumulative percentage is plotted in another axis. Pareto chart is useful in answering what areas have the greatest impact? Which steps in the process create the majority of errors? However, 80/20 rules may not whole in all situations. In general, if the vast majority of the problems are not attributable to a minority of the categories or if the categories take turns “leading the parade”, then it will be a mistake to tackle the critical few categories.

Bar Charts:

• You may use bar charts to compare values or amounts across different categories
• One scale is a list of categories
• The other scale shows the possible values for some measure
• Each bar displays the value of the measure for a particular category
• The lengths of the bars make the comparisons between categories

Pareto Charts:

• Categories are organized according to the height of the bars, except that a miscellaneous category is placed last
• Costs, rather than counts, are preferred measure for use in a Pareto Chart
• A cumulative percentage curve adds perspective to the char and allows you to check to see if the Pareto principle is present.
• The stair-step of the bars and the cumulative percentage curve make the comparisons

Histogram

• You may consider using a histogram whenever you want to see how the values from some measure vary
• One scale consists of possible values for the observed measure
• The other scale consists of frequency values
• Each bar on the histogram shows the frequency with which the observed measure falls into the relevant interval
• The bars show the mounding of the data and the taper-off toward each extreme

Running Records/ Time-series

• You may use a time-order sequence of value to visualize the process
• Unlike other charts which are fixed in a single point in time, the time component is varying in running records
• One scale consists of possible values for the observed measure
• The other scale consists of the time period
• The values of this one measure are plotted in their time-order sequence
• The variation of this measure over time is displayed, and the comparison is a comparison of a single measure with itself at different time periods.

Pie Charts

• Similar to the function of bar charts but the amount is converted into percentages.

This chapter is generally a quick read if you already know those charts.