A histogram is a graphical representation of a frequency distribution. The data is divided into class intervals and denoted by rectangles. The rectangles are made on the X axis. On the Y axis, the analyst plots the frequencies of the data. Each rectangle represents the numbers of frequencies that lie within that particular class interval.

Analyze the histogram to see whether it represents a normal distribution. Once you have plotted all the frequencies on the histogram, your histogram would show a shape. If the shape looks like a bell curve, it would mean that the frequencies are equally distributed. The histogram would have a peak. The peak represents the highest values of the data. In this kind of distribution, both sides of the peak would have almost equal numbers of data frequencies. For example, if a company is using the histogram to understand the preferences of customers across two different choices, a normal distribution would represent that majority of the customers are indifferent.

Analyze the histogram to see whether it represents a skewed distribution. A skewed distribution histogram is one that is asymmetrical in shape. All the frequencies lie on one side of the histogram. The distributions lie on either the right-hand side or the left-hand side of the peak. Through this diagram, the analyst knows which side of the histogram he must concentrate on.
For example, if the company is studying the customers’ tolerances to price changes, with this type of histogram the company would see the price changes that are most acceptable.

Analyze the histogram to see whether it represents a bi-modal distribution. In these kinds of histograms there are two peak points. These points represent the highest values. For example, the company may be assessing the workers’ productivity levels across different hours in the day. The examination may reveal that the workers are most productive at 9 a.m. and 4 p.m. Therefore, there would be two peaks in the histogram.

Analyze the histogram to see if it represents a truncated distribution. The histogram of a truncated distribution looks very much like a normal distribution histogram with its edges cut off. For example, the company may be running quality checks on the raw material inventories and there may be no figures in the extreme limits.