ahmadbelb / manf

Purpose

To provide analysis tools and metrics useful in manufacturing environments.

I am slowly generating the documentation and, as that is maturing, I will begin to move information from this readme.md into that location. If you don't find something here, head over to the documentation.

Project Maturity

Every effort is being made to ensure that the results are accurate, but the user is ultimately responsible for any resulting analysis.

The API should not be considered stable until v1.0 or greater. Until then, breaking changes may be released as different API options are explored.

During the v0.X.X versioning, I am using the package in my own analyses in order to find any bugs. Once I am reasonably satisfied that the package is feature complete, usable, and bug-free, I will break out the v1.X.X releases.

Usage

Visualizations with Jupyter Notebooks

Visualizations work approximately as expected within a jupyter notebook.

data = np.random.normal(0, 1, size=30)  # generate some data
manufacturing.ppk_plot(data, lower_control_limit=-2, upper_control_limit=2)

There is a sample jupyter notebook in the examples directory.

Cpk Visualization

The most useful feature of the manufacturing package is the visualization of Cpk. As hinted previously, the ppk_plot() function is the primary method for display of Cpk visual information. First, get your data into a list, numpy.array, or pandas.Series; then supply that data, along with the lower_control_limit and upper_control_limit into the ppk_plot() function.

manufacturing.ppk_plot(data, lower_control_limit=-2, upper_control_limit=2)

In this example, it appears that the manufacturing processes are not up to the task of making consistent product within the specified limits.

Zone Control Visualization

Another useful feature is the zone control visualization.

manufacturing.control_plot(data, lower_control_limit=-7.0, upper_control_limit=7.0)

Features Map

Analysis

• Ppk analysis
• Ppk plots/histograms
• Cpk analysis/plot/histogram by subgroup
• In-control/out-of-control analysis (do Ppk and Cpk converge to approximately the same value)
• Control chart plot (see Control Chart Rules)
  • Beyond limits violations highlighted (one or more points beyond the control limits)
  • Zone A violations highlighted (2 out of 3 consecutive points in zone A or beyond)
  • Zone B violations highlighted (4 out of 5 consecutive points in zone B or beyond)
  • Zone C violations highlighted (7 or more consecutive points on one side of the average - in zone C or beyond)
  • Trend violations highlighted (7 consecutive points trending up or down)
  • Mixture violations highlighted (8 consecutive points with none in zone C)
  • Stratification violations highlighted (15 consecutive points in zone C)
  • Over-control violations highlighted (14 consecutive points alternating up and down)
• Gage R&R analysis

Usability

• Import from CSV
• Import from MS Excel
• Create documentation using sphinx
• Generate reports

Gallery

Currently, no distinction is made between Ppk and Cpk, so the entire chart shows the Cpk.

Some of the data for the zone control chard was manipulated in order to display the results. Note that, if a phenomenon is not present within the data, then it will not be plotted at all.