Operation chart, also known as chain chart, is a special scatter chart, which shows the changes of any measurement characteristics with time. The purpose of analyzing the operation diagram is to confirm whether the wave pattern is caused by general factors or special factors. Operation chart can be used for graphic analysis of any data organized in time series and measured in continuous scale. Variation will occur in the process, and variation caused by routine reasons is a normal part of the process; Another type of change (called special cause) comes from outside the system and leads to recognizable patterns, deviations or trends in the data. The operation chart shows the process affected by special reasons. The process is controlled when only conventional reasons affect the process output. Running charts perform two kinds of randomness tests, providing information about non-random changes caused by trends, vibration, mixing and aggregation.

The drawing method of operation diagram is as follows:

1. Draw a line chart of data in chronological order.

2. Find the median of the data, draw a horizontal line on the graph and mark this line with "median".

Operation chart can be used to judge whether the process is affected by special factors. There are several ways to judge:

The length of the run refers to the number of consecutive points on the same side of the median. Unless the process is affected by abnormal factors, it is unlikely that a long series of continuous points will fall on the same side of the median. Therefore, checking the length of the run becomes a method to test the special fluctuation factors.

However, some values are exactly equal to the median. If only one point falls on the center line, ignore this point. In the case of odd data points, usually at least one data value is equal to the median. If more than one point falls on the center line, distribute these points to both sides according to 50% of each side. If the length of the longest run is long, this process is likely to be affected by special fluctuations. Whether the run length is abnormal or not should be judged by statistical test.

3. Trends

There should be no abnormal sequence of continuous rise and continuous fall in the operation chart. If this happens, it implies an abnormal trend. If there are many points rising or falling continuously, there may be process deviation caused by special factors, and it is necessary to judge whether the trend in the operation chart is abnormal through statistical test.