Time Series Analysis

Contains tools for analyzing time-series data.

Command-Line Tools

The analyze_series command is meant to be used in the command-line in typical *nix fashion, expecting:

• one number per line
• input on standard input or as a file argument to analyze_series

analyze_series will analyze the points in the provided series and print the following descriptive statistics:

• median
• mean, sample standard deviation
• upper and lower control limits
• points and associated values falling outside of control limits

analyze_series can be used to help implement a run-by-run control program using Control Charts. analyze_series currently supports 'Rule 1' testing for points that are +/- 3-sigma from the series' mean. Support for detecting suspicious runs above/below the mean is planned (issue #3).

Usage

execute analyze_series -h for full usage details.

excerpt:

usage: analyze_series [-h] [-v] [--assert-last-point-in-control] [input_file]

positional arguments:
  input_file            input file to process, defaults to standard input

optional arguments:
  -h, --help            show this help message and exit
  -v, --verbose         print additional debug information
  --assert-last-point-in-control
                        assert the last point in the series is in-control and
                        exit with status code '2' if last point is not in
                        control

Examples

Analyze a file on the filesystem:

$ ./analyze_series resources/examples/multiple_control_limit_violation.txt
median: 0.0
mean: 0.0357142857143
std dev: 1.41375432756
lower control limit: -4.20554869697
upper control limit: 4.2769772684
points outside of lcl: [(3, -5.0)]
points outside of ucl: [(28, 7.0)]

Analyze data arriving on standard input:

$ cat resources/examples/under_control.txt | ./analyze_series
median: 10.05
mean: 10.05
std dev: 0.873689494805
lower control limit: 7.42893151558
upper control limit: 12.6710684844
points outside of lcl: None
points outside of ucl: None

Analyze data and return error if last point is not in control:

$ cat resources/examples/upper_control_limit_violation.txt | ./analyze_series --assert-last-point-in-control; echo "exit status: $?"
median: 0.0
mean: 0.142857142857
std dev: 1.11269728053
lower control limit: -3.19523469873
upper control limit: 3.48094898444
points outside of lcl: None
points outside of ucl: [(27, 4.0)]
stderr: last point (index=27, value=4.0) is out of control
exit status: 1

Acknowledgement and Thanks

The time-series-analysis project builds-upon and is made-easy by the wonderful work of:

• Python
• NumPy - the fundamental package for scientific computing with Python
• pandas - provides high-performance, easy-to-use data structures and data analysis tools for Python

Thank you!