KPI Analysis Introduction¶
You can add an analysis to an advanced KPI in the same window in which you create the KPI. You can also open a previously saved advanced KPI and add an analysis to it via the Edit Advanced KPI page. When you run a KPI analysis, the system opens an additional browser window where it displays the analysis results in a chart and results table. Results tables can display up to 2,000 rows of data. Results are stored in the default Output location provided by the system for Integrated Data Lake data sources. You can open KPI analyses results in an Operations Insight (OI) dashboard using the KPI Analysis widget.
KPI Analysis features include¶
Monitoring: a two-dimensional analysis of KPI results charted against a variable, such as the time dimension or another field in the data. This type of analysis is used for Integrated Data Lake data sources, and helps users monitor KPI values for a specific field/variable. Users can switch the display for a monitoring analysis back and forth between bar and line charts.
Discovery: a multi-dimensional analysis that displays a bubble chart. Bubble size and placement represent the distribution of KPI values against contributing factors. This type of analysis helps users discover data trends and outliers among the KPI analysis data. Discovery analyses can contain up to seven Group Results By settings. The bubbles in the discovery analysis chart display the value for the field relative to the whole, that is, by weighted percentage; however, the data in the results table is rendered in values, rather than wieghted percentage.
Data Drill Down: a feature of the Analysis Results table that allows you to drill down into the data for any of the fields contained in the analysis results. Group by filters applied in the analysis carry into this drill down.
Combination History: a feature of the Analysis Results table that lets you examine the history of the various field combinations contained in the analysis results. the discovery analysis feature that lets you drill down into the analysis data from the analysis results table to analyze the impact of various combinations of the analysis fields. Group by filters applied in the analysis carry into this drill down.
About Configuring Analysis Fields in General¶
Here are some important points to bear in mind when working with KPI Analysis configurations:
- All analysis settings are optional except the first Group Results By setting.
- If you don't set a Group Results By setting, the system will distribute KPI analysis results along the X-Axis by the default grouping, which is by date with month as the time interval for Integrated Data Lake data sources, or by minute with 2 minutes as the time interval for time series data sources.
- Analysis settings are independent; thus, configuring one area, such as adding a schedule, does not require you to configure another area, with two exceptions:
- Notification: if you want to be notified when analysis results cross a threshold, you must configure a threshold.
- Autodiscovery: to add autodiscovery, a threshold is required.
Analysis Settings Description¶
Here is a brief description of the settings for an analysis:
Group Results by with Integrated Data Lake (IDL) sources: when using IDL data sources, you can group results by date or field and can add a minimum of one and up to a maximum of seven groupings. Adding one or two Group Results By settings produces a monitoring analysis. Adding three Group Results By settings (and up to a maximum of seven) produces a discovery analysis.
Group Results by with Time Series data sources: when using time series data sources you can add only one Group Results By setting by selecting a time interval and specifying a number; for example, you could select an interval of 'Minute' and specify a number from 1-60, or 'Hour' and a number from 1 -24, or 'Day', and 1 day. The time series configuration produces a monitoring analysis.
Filters: you can add filters to both dates and fields. Date filters can be fixed dates or rolling dates, and you can narrow analysis data by adding a filter to any field in the data. Note: only filters that have been configured with the Group Results By setting display in the Variables drop-down list.
Schedule: schedule the KPI analysis to run automatically at the date and time you specify.
Share and Notify: set up a threshold that, when crossed by analysis results, notifies you and those you share the notification option with. The share feature applies to the notification only, and does not mean you have shared the KPI or KPI analysis with others.
Autodiscovery: add autodiscovery to a KPI analysis to run when a threshold is crossed.
Any questions left?
Except where otherwise noted, content on this site is licensed under the MindSphere Development License Agreement.