top of page

# Frequency Plots

Frequency plots are a visual tool used to analyze numerical data by showing the pattern of how often different values occur. Frequency plots provide us with a picture of the shape, range, and center of the data, as well as where the data falls relative to targets or specifications. The most common types of frequency plots are histograms and dot plots. Histograms use bars to represent how many times a value (or range of values) has occurred.

Histogram:

Dot plots use x’s or dots to show when a value has occurred.

Dot Plot:

The shape of a frequency plot helps to determine what strategy to follow to make improvements to a process. There are many different shapes which indicate various strategies for improvement, but we will focus on four shapes today: bell curve, skewed, bimodal, and outliers.

Bell Curve: A bell curve indicates a normal distribution.

Example Bell Curve:

In the case of a bell curve, you would follow the DMAIC process to make improvements.

Skewed: A skewed distribution is a common shape for a frequency plot and occurs when the data is asymmetric with a long tail on one side. A positive skew is common when measuring values that cannot be less than 0, such as number of errors, wait time, etc. A negative skew occurs when there are more data points near the upper end of the scale, such as retirement age.

Example Positive Skew (long tail):

The most common way to analyze data which shows a skewed is to perform a data transformation.

Bimodal: a bimodal frequency plot shows two distinct peaks in the data. Usually, when this occurs, there are two different groups contributing to the data.

Example: Bimodal

In order to analyze this data, you should attempt to stratify your data to identify what the different groups are so that you can better understand the overall process.

Example: Bimodal Stratified

Based on this stratification, you can focus on why the two sites are so different and maybe transfer some best practices from Site A to Site B.

Outliers: occur when a few points differ significantly from the rest of the data.

Example: Outliers

If you notice outliers on a frequency plot, investigate these specific points instead of focusing on the overall process.