Using a time series plot with six sigma

Six Sigma is a business management strategy used to help remove the causes of defects and errors in both manufacturing and business processes. The Six Sigma strategy asserts that continuous efforts to achieve stable and predictable process results are vital to business success and the manufacturing and business processes lend themselves to having characteristics that can be easily measured, analyzed, improved, and controlled.

It was originally developed by Motorola and follows a strict sequence of steps measured by a financial marker (cost reduction or profit increase). The process itself utilizes a set of quality management methods, statistical methods, and a special infrastructure of people who are expert in these methods. These experts are referred to as "Black Belts".The Six Sigma strategy is data-driven and uses DMAIC and DMADV sub-methodologies.DMAIC stands for define, measure, analyze, improve, and control. DMADV stands for define, measure, analyze, design, and verify. A problem or concern area is identified and objectively defined. The output of that process is then measured and analyzed for variations. The process can be improved if the variations are eliminated.

In statistics, time series is a method of gathering a sequence of data points at successive times spaced at usually consistent intervals. The analysis of time series data is primarily used to identify a problem. A time series plot is a chart that graphically shows a set of observed data taken at varying points in time. There are two different types of Time Series data.

Continuous: This data is derived from a process that requires a continuous stream of outputs. An example would be a lie detector. The data provided cannot be taken at a snapshot, but continuously.

Discrete: This data is collected at regularly spaced intervals. This could be on the hour or every 2 weeks.

Time series graphs show outputs as a function of time. By doing so it is easy to identify the nature of the process. When doing the analysis you should be mindful of outliers, discontinuities, trends, periodicities, and time intervals.

Outliers are values that do not seem to be consistent with the rest of the data. Graphically, an outlier will be a point further away from all the rest. It is out of the ordinary. Outliers can signal areas that may need to be addressed.

Discontinuities are gaps or breaks in the data that should be continuous. Graphically you might notice a large gap or spacing between data points. Again, these can flag areas that may need to be retooled or addressed.

Trends show a general tendency and direction. These give insight into the overall process and can provide great insight in where to begin the refinement process.

Periodicities are recurrences at regularly spaced intervals. Graphically these might be represented by data trends that are identical every six months on your time series.

Make sure you don't overlook the time intervals that are specific to the process you are working with. Some time series data collected may include previous and future periods that are not pertinent to the process being addressed.

Observations collected and charted should be done in order. This is important in that the observations may be co-dependent. Scatter plots charts are most commonly used for time series data observations. The vertical axis becomes your X value while the horizontal is your T (time) value.

Time Series plots are useful with Six Sigma because they allow you to see the trends, peaks, valleys and averages within a time period. You want and need to be able to analyze and identify the underlying theory of the points like what generated them, where they came from, and for making additional forecasts and predictions. They are important for gaining these insights.