Creating a missing data strategy for a data set
Setting up a missing data strategy for a specific data set rather than as a global strategy provides a way to fine-tune your
analytic results.
Prerequisites: All points are tagged in preparation to run an analytic query.
- Expand and double-click Definitions.The Analytic Data Manager view opens.
- To create a missing-data definition, click New.The New window opens.
- To accept the default
Type to Add(Analytic Data Definition), click OK.A second New window opens.
- Name the strategy, enter a tag in the
Idproperty, enableUse This Value(set totrue), select anInterpolation Algorithm, define theRawDataFilterproperties used to evaluate each record and click OK.The Id tag identifies the data set to include all points tagged in this example with mt:missingData. Setting theLow Limiton theRawDataFilterto 1.00 configures the strategy to remove any value that is less than 1, which removes any zeros from the data set. - Add an AnalyticWebChart to a PX view.The reason to use this type of chart is because it allows you to define a tag that identifies the data to return on the chart.
- Double-click the chart to open the properties window.The widget opens.
The
dataproperty should already be set to the tag you entered when you created the definition. This capture shows the same binding configuration but with KNN-1 enabled. - Configure the
dataproperty to the tag you entered when you created the definition. - Set any other properties you ignored when you created the definition, including
aggregation,rollup,missingDataStrategy, or outlier handling. and click OK.You should be able to identify the interpolated date points in the chart.
The tool tips identify the interpolated data.
- For another way to configure each chart, click the configuration icon (
).The Settings window opens. In this window changes to theMissing Data Strategyoverrides directly in the chart. An end user could use this configuration feature.