Which Charts Do I Use When Analyzing Data?

Lean Six Sigma black belts are trained to analyze data and look for anomalies and trends. The things that they look for are the 7 deadly wastes. These wastes are overproduction, waiting, inventory, extra production, transportation, (non value added) motion, and defects.

Part of analyzing this data requires a basic understanding of processes and data, and how these can be displayed. It is often easier to analyze trends and anomalies in picture form. Graphs, charts and flow charts depict a visual summary of data elements and relationships. When thinking about what charts to use, think about the essence of relationships and what you are trying to show.

How do you know what kind of graph or chart to use?

Are you looking for trends, year over year, or month over month such as inventory turns, inventory dollars or the number of defects per product line?

Bar charts

If so, you will want to first come up with either an average or a total sum for the time period, to be able to show aggregates in categories of information such as a date series, and then compare summarized trends. Use a bar graph to show time periods and comparisons between months or years after creating a distribution table to categorize the data.

Pie charts

If you want to compare a percentage of a whole, such as the number of orders by customer to one another, you have to create a ratio by comparing the total to each portion. Use a pie chart to visually display the pieces in relationship to the whole. For example, let's say you have 5 customers. You provide product or services to these customers over the period of a year. Who is your best customer? Break down total sales by customer first, and then come up with a grand total. Divide each of the customer sales into the total sales to find the percentage related to that customer.

Note that you may have one customer that gives you one order a year that is worth double the amount of another customer's 10 monthly orders. Be cognizant to measure dollars in this situation, rather than number of units sold.

Flow charts

If you are you looking for areas of overproduction, waiting, inventory, transportation, or movement then you will want to use flow charts or spaghetti charts.

Lean experts often start their analysis of a project by mapping out the process. A flow chart is simple, showing when a process starts, what happens in the process, and how it ends. A value stream map is more complex. It shows chronological movement between people, operations and communications indicating parallel, serial and backwards flows (rework).

Value stream map and lead time ladders

To use a value stream map effectively, you have to determine the boundaries of the process that you are trying to visualize. A flow map will contain macro steps that break off into different paths at decision points but it is usually limited to one operation, or step in the process. A value stream map on the other hand will cross functional group boundaries by showing the steps it takes to process a product or service from cradle-to-grave, or from the very beginning of the process to the very end.

At the bottom of a value stream map when using Six Sigma Lean methods, you will have a lead time ladder. These measurements break out the time it takes to wait for a process to start (queue time), the time it takes a process to be set up (set up time) the time it takes to do the process (processing time) and it will also break out time wasted (time in queue not moving, requesting confirmation or approval or more information (rework or inspection). In addition to the flow, the communication and the times, it also shows where inventory stops.

Stomping out waste

Whether inventory is a piece of writing that needs to be edited to become published, or it is a physical product, or it is human intellect sitting around waiting, every time inventory stops moving, it costs money.

These are bottleneck areas that get analyzed further with trend charts and histograms. Histograms in Excel are created to show a specific set of data points. These data points can then be compared or correlated to other similar data points to show normal and abnormal processing. The abnormal spots are analyzed for improvement first. Then energy is spent looking at the overall process and removing waste, improving process flow and communications.

When trying to figure out whether to use a chart, a graph, or a flow map, think about the relationships to the data elements and what you are trying to show, then pick the right graph, chart or flow map to show the right data.

Sources:

http://exceluser.com/intro/three-report-mistakes.htm

http://www.vertex42.com/ExcelTemplates/fishbone-diagram.html

http://www.leaninnovations.ca/seven_types.html

http://www.dspguide.com/ch7/3.htm