# Making a Tableau Control Chart to PinPoint Outliers in Data

A** Tableauâ€¯control chart**â€¯is a graph used to study how a process changes over time. All processes have some variability. Thatâ€™s normal. But large shifts or swings are cause for study and indicate something has changed about the way your process is behaving. They are used to pinpoint sources of variation.

Data are plotted in time order. Aâ€¯control chartâ€¯always has a central line for the average, an upper line for the upper controlâ€¯limit, and a lower line for the lowerâ€¯control limit. These lines are determined from historical data and typically are based on standard deviations from the average or median line in the center.

If the process is in control (and the process statistic is normally distributed, which is likely), 99.73% of all the points will fall between the control limits (usually 3 standard deviations above and below the average). Any observations outside the limits, or systematic patterns within, suggest the introduction of a new (and likely unanticipated) source of variation, known as aâ€¯special-causeâ€¯variation.

You are looking for any point above the control limit lines, or a run of 7 points that all fall either above or below the average/median line, or an upward or downward trend of 7 values. For more info, see: https://public.tableau.com/s/blog/2013/11/how-make-control-charts-tableau

### History of Control Charts

Using control charts is a great way to find out whether data collected over time has any *statistically significant* signals, or whether the variation in the data is merely noise.

Control charts were invented and first used by Walter Shewhart from Western Electric Company in the 1920s. It was specifically made in the context of industrial quality control.

However, the “six sigma” movement brought control charts to life, into mainstream data analysis use. In an attempt to reduce process variation, they used it measure process behavior. In theory, the less process variations, the fewer defects there will be. As a result, they improve overall quality of processes.

With a goal if reducing variations, process specialists have to look into historical data and figure out if there are any source of special-cause variation or signals in it.

### Anatomy of a control chart

In order to understand Control charts better, let us dissect and break down the elements of the control chart. Then, we will learn how to make one step-by-step.

You can find the following elements in a control chart:

- X and Y Axis or the time series data
- The average line
- The control limits
- Lower Control Limit – LCL
- Upper Control Limit – UCL

- Signals:
- Outliers – data points that are above the UCL or below the LCL
- Trends – 6 or more points that are either entirely ascending or entirely descending
- Shifts – 9 or more points either entirely above or entirely below the average line

### How to Create Tableau Control Charts

Now that we’ve got the basics covered, and you have a basic understanding of the Tableau Control Chart Anatomy, let’s see how it’s done.

- Create a line chart.

- Add a reference line, by right clicking on the axis or on the Analytics pane at the left, and set it to be a Line, Per Pane, and set the value to Average (or Median if desired).Â

- Right click on the axis again and add another reference line.

- Make this one a Distribution. Set it to Standard Deviation. Type in -3,3. If that seems too wide, you can also choose either -1,1 or -2,2.

- Click ok and look for any outliers in your data.

A slightly more advanced version with color will look like this:

To sum it up, a simple control chart will allow you to find outliers but wouldn’t it be better if you make it the outliers more obvious? We can teach you exactly how to do it in our Tableau training courses!