Laserfiche WebLink
<br />9 Quality Control <br />Version 10.00.01, 2010-05-16 <br /> <br /> <br />This chapter discusses how TSTool software is quali ty controlled and how to use TSTool for quality <br />control. <br /> <br />9.1 Using TSTool to Quality Control Data <br /> <br />TSTool can be used to perform quality control on time series data. Two primary commands are: <br /> <br />1. CheckTimeSeries() – this command checks individual time series values for out of range, <br />missing, greater than, etc. Values that are det ected can be flagged with a string and optionally <br />can be set to missing. The HTML summary (see WriteSummary() command) will indicate <br />flagged values and graph products can label data points with flags. Flags can be written to data <br />management systems if flags are supported. <br />2. CheckTimeSeriesStatistic() – this command calculates a statistic (e.g., count of data <br />values greater than a criteria) and then checks the statistic against a criteria (e.g., is count greater <br />than a criteria). <br /> <br />The above commands can be used to check for out of range and other unusual da ta conditions. If values <br />are replaced with missing, the missing values can be filled using the fill commands. Additional quality <br />control can be performed by using a combin ation of other commands. For example, the <br />NewStatisticYearTS() command computes an annual statistic time series, which can then be <br />checked using one of the above commands. The Cumulate() command also is useful for visualizing <br />trends. <br /> <br />9.2 Using TSTool to Quality Control System Functionality <br /> <br />The previous section described how TSTool can be used to quality control data. Another aspect of quality <br />control is ensuring that a system is working. Fo r example, TSTool accesses data from many input sources <br />and it may be difficult for the maintainers of data systems to know if systems are running (is the system <br />up?) and that data are being properly returned (are values reasonable?). TSTool can be used as a test <br />runner to check for system uptime, for example: <br /> <br />1. Blanket test. Generate a command file to read a representative number of time series (or all) from <br />a system. It may be necessary to set the input period to be short. If any errors occur, they need <br />to be resolved. <br />2. Data count test. Generate a command file to read a representative number of time series. Specify <br />the input period as an appropriate value (e .g., short for real-time data). Use the <br />CheckTimeSeriesStatistic() command to generate a count of data values in the period <br />and check for reasonable values. <br /> <br />Other tests also can be performed. The following sec tion describes how very specific tests can be run to <br />test software functionality. A similar approach can be used to test systems, if it is known that expected <br />results from the system will not change over time. <br /> <br /> Quality Control - 1 117