This glossary summarizes terms found in the DeepSee documentation. If you have not yet done so, InterSystems highly recommends that you read “Basic Concepts” in Defining DeepSee Models.
An operation that a user can start by using a control (such as a button) on a dashboard. DeepSee provides a set of standard actions (such as applying a filter, navigating to another dashboard, and others), and you can add custom actions. See “Defining Custom Actions” in the DeepSee Implementation Guide.
An age dimension is a dimension that contains age levels. An age level groups data by an age, relative to the cube build time, computed from a date or time value in the source data. Age dimensions and age levels are not generally recommended, because they require nightly rebuilds.
The All level is a special, optional level, which appears in all the hierarchies of a dimension. If defined, this level contains one member, the All member, which corresponds to all records in the cube. You can use the All member to create a summary line in a pivot table.
Business Intelligence, a set of tools and techniques that transform raw data into insights that can improve the operation of a business or other organization. BI is intended to support a measurement-based approach to making strategic and tactical decisions.
For details, see “Compiling and Building Cubes” in Defining DeepSee Models and “Keeping the Cubes Current” in the DeepSee Implementation Guide
A two-dimensional array of data generated by a running Ensemble production and generally providing data relevant to or about that production. Like pivot tables, business metrics can be displayed on a dashboard, within a widget. For information on creating Ensemble business metrics, see Developing Ensemble Productions.
A Ensemble concept that allows nontechnical users to change the behavior of Ensemble business processes. You can use them in source expressions in DeepSee cubes; see “Details for Source Expressions” in Defining DeepSee Models. For details on Ensemble business rules, see Using Business Rules with Ensemble.
A measure that is based on other measures via an MDX expression. The phrase calculated measure is not standard in MDX, but this documentation uses it for brevity. Formally, a calculated measure is a calculated member that belongs to the Measures dimension.
A member that is based on other members via an MDX expression. You can define two kinds of calculated members:
A calculated measure is a measure is based on other measures. (In MDX, each measure is a member of the Measures dimension.)
For example, one measure might be defined as a second measure divided by a third measure.
The phrase calculated measure is not standard in MDX, but this documentation uses it for brevity.
A non-measure calculated member typically aggregates together other non-measure members. Like other non-measure members, this calculated member is a group of records in the fact table.
See “Defining Calculated Members” in Defining DeepSee Models.
A special kind of subject area that combines multiple cube definitions (typically two) and that enables you to create pivot tables that contain elements from multiple cubes. See “Defining Shared Dimensions and Compound Cubes” in the Advanced DeepSee Modeling Guide.
A special kind of DeepSee dimension whose members are computed at runtime via an SQL or MDX expression. See “Defining Computed Dimensions” in the Advanced DeepSee Modeling Guide.
Computed dimensions do not have any association with calculated members. A computed dimension is specific to DeepSee. A calculated member is a standard concept in MDX.
A class that extends %DeepSee.UserLibrary.Container. This class can contain the definitions of pivot tables, dashboards, and other DeepSee folder items. When you compile this class, Caché generates those folder items, replacing any current definitions that they might have. See the DeepSee Implementation Guide.
An interactive element on a dashboard. Controls include drop-down lists and buttons.
An model of your data that defines elements that can be used in MDX queries. These elements determine how you can query your data, specifically, a set of specific records (such as patient records or transaction records). The set of records is determined by the source class for the cube. For an introduction, see “Basic Concepts” in Defining DeepSee Models.
A mechanism in DeepSee that enables you to define multiple similar cubes. This mechanism has no relationship to class inheritance. See “Using Cube Inheritance” in the Advanced DeepSee Modeling Guide.
A listing, specifically one of the following special kinds of listings:
A listing that uses a custom SQL query that retrieves fields from some other table, not the source table used by the cube, and not a data connector. See “Defining Listings” in Defining DeepSee Models.
A listing that consists of listing fields chosen by the user, in the Analyzer. See “Performing Ad Hoc Analysis” in Using the DeepSee Analyzer.
An interactive display of data, particularly data that provides a high-level data of a business. See Creating DeepSee Dashboards.
A class that extends %DeepSee.DataConnector. A data connector maps the results of an arbitrary SQL query into an object that can be used as the source of a cube. Typically, a data connector accesses external non-Caché data, but you can also use it to specify an SQL query against Caché, including an SQL query on a view. See “Defining and Using Data Connectors” in the DeepSee Implementation Guide.
A container for levels. A dimension contains one or more hierarchies, which in turn contain levels. For example, a single dimension might contain multiple hierarchies related to allergies. There is no formal relationship between two different hierarchies or between the levels of one hierarchy and the levels of another hierarchy. The practical purpose of a dimension is to define the default behavior of the levels that it contains — specifically the All level.
See “Defining Dimensions, Hierarchies, and Levels” in Defining DeepSee Models.
The table in which DeepSee stores the members of a level and any properties they have. See “Details for the Fact and Dimension Tables” in Defining DeepSee Models.
Examine a row of a pivot table and see the data for that row displayed in a more granular way. For example, a row might display data for a year, and you would drill down to see data for that year, broken out by month. DeepSee supports multiple forms of drill down. See “Performing Ad Hoc Analysis” in Using the DeepSee Analyzer.
Informally (although not in this documentation), the phrases drill down and drill through are sometimes used interchangeably, and it is wise to double-check which phrase is intended.
Formally, to drill through means to display a listing. Internally, DeepSee uses the MDX DRILLTHROUGH statement when it displays a listing. See “Performing Ad Hoc Analysis” in Using the DeepSee Analyzer.
Informally (although not in this documentation), the phrases drill through and drill down are sometimes used interchangeably, and it is wise to double-check which phrase is intended.
An expression (<expression> element) whose value is available while DeepSee is building a row in the fact table. You can define an expression that uses complex or time-consuming logic, and then you can base multiple cube elements on the expression. Expressions are for use during cube build only and are provided for efficiency.
See “Other Options” in the Advanced DeepSee Modeling Guide.
A row in the fact table.
A generated structure that DeepSee queries directly. When you compile a cube definition, DeepSee generates a fact table class. When you build a cube, DeepSee creates records for this table and indexes them. See “Basic Concepts” in Defining DeepSee Models.
A restriction on the data. DeepSee provides two simple ways to filter data: member-based filters and measure-based filters. You can combine these, and more complex filters are also possible, especially if you write MDX queries directly. For an introduction, see “Filters” in “Basic Concepts” in Defining DeepSee Models.
Any of the following DeepSee items:
DeepSee folder items are visible in the Studio Workspace window, where they are shown in the Other folder.
See map listing.
An organization of levels. Levels belong to hierarchies (which belong to dimensions). A hierarchy can contain only single level or can contain multiple levels. If it contains multiple levels, the “higher” levels of the hierarchy are less granular then the “lower” levels. That is, each member of a higher level contains a larger set of records than does a member of a lower level.
In casual usage, a higher level is called the parent of the lower level. However, it is useful to remember that the hierarchy is a actually a hierarchy among members. Thus it is more accurate to state that a member of the higher level is the parent of one or more members of the lower level. Conversely, any member of a lower level is the child of exactly one member of the higher level.
Hierarchies provide additional features beyond those provided by levels; see “Hierarchies and Dimensions” in Defining DeepSee Models. Also see “Defining Dimensions, Hierarchies, and Levels” in the same book.
A special kind of dimension that analyzes an iKnow measure, which in turn is a measure based on unstructured text. See “Using Unstructured Data in Cubes (iKnow)” in the Advanced DeepSee Modeling Guide.
A special kind of measure that is based on unstructured text. You cannot display iKnow measures directly in pivot tables. Their purpose is to provide data for use by iKnow dimensions. See “Using Unstructured Data in Cubes (iKnow)” in the Advanced DeepSee Modeling Guide.
A class based on %DeepSee.KPI. In most cases, a KPI uses a query and displays a result set. Like pivot tables, KPIs can be displayed on a dashboard, within a widget. You can also use KPIs as building blocks for calculated members (including calculated measures). See “Defining Basic KPIs” and the chapters that follow it, in the Advanced DeepSee Modeling Guide.
A cube element that enables you to group records. A level consists of members, each of which is a set of records. See “Basic Concepts” in Defining DeepSee Models. Also see “Details of Defining Levels” in the same book.
A level that is based upon a list value. For example, a patient can have multiple diagnoses. The Diagnoses level groups patients by diagnosis. With a list level, it is possible for a given record of the source class to have multiple values and thus for that record to belong to multiple members of the level.
An SQL query that accesses the lowest-level records associated with one or more cells of a pivot table. See “Defining Listings” in Defining DeepSee Models.
A class that defines a group of listings. Listing groups are created in the Listing Group Manager. The purpose of this tool is to enable you (and your customers, if appropriate) to define listings outside of cube definitions and without needing access to the Architect. See “Defining Listing Groups” in Defining DeepSee Models.
A <listingField> element defined in a cube definition. Users can select the listing fields to include, when they create custom listings in the Analyzer. See “Defining Listing Fields” in Defining DeepSee Models.
This phrase can also refer more generally to any field in any listing.
A listing that contains location data and is displayed as a map. Each pin on the map corresponds to a source record.
Note that in order to use the Google Maps API, you must obtain an API key. See Specifying Basic Settings in the DeepSee Implementation Guide for more information.
Note that in order to use the Google Maps API, you must obtain an API key. See Specifying Basic Settings in the DeepSee Implementation Guide for more information.
A cube element that (with rare exceptions) aggregates values across multiple records. Each measure is based on a source value, which is either a class property or an ObjectScript expression. The definition of a measure also includes an aggregation function, which specifies how to aggregate values for this measure. See “Basic Concepts” in Defining DeepSee Models. Also see “Defining Measures” in the same book.
A set of records. Every level has one or members. See “Basic Concepts” in Defining DeepSee Models. Also see “Details of Defining Levels” in the same book.
An interactive, drillable display of data, generally with rows and columns, designed for specific user roles or for specific areas of your user interface. A pivot table is based on an MDX query that is executed at runtime can respond to input such as filter selections made by the user. Internally it obtains values from a cube. See Using the DeepSee Analyzer.
An element that is intended to be used in pivot tables, specifically, in selected parts of the query that defines the pivot table. When a dashboard displays the pivot table, that dashboard can include a control with which the user can change the value of the corresponding pivot variable. See “Defining and Using Pivot Variables” in Using the DeepSee Analyzer.
Pivot variables are entirely different from runtime variables.
A specialized form of KPI that defines one or more computations to use in the Analyzer and in queries. Plugins are especially appropriate for complex or time-consuming computations. For example, you might have a computation that uses several different parts of the source record, as well as external information; a plugin would be suitable in this case. See “Defining Plugins” in the Advanced DeepSee Modeling Guide.
A value that is specific to a member of a given level. If a level has a property, then each member of that level has a value for that property; other levels do not have values for the property. You can use properties in queries in much the same way that you use measures. In DeepSee, you can also use properties for other purposes such as controlling member names and controlling the order in which member are sorted. See “Defining Properties” in Defining DeepSee Models.
A quality measure is similar to a calculated measure because it is defined by a formula that combines MDX expressions. You specify the subject area or subject areas in which it is available, and you can control whether the quality measure is published (and thus available in the Analyzer). Each quality measure is a subclass of %DeepSee.QualityMeasure.QualityMeasure.
For information, see “Defining Quality Measures” in the Advanced DeepSee Modeling Guide.
A connection between two cubes that makes the dimensions of one cube available in the other cube (and possibly vice versa). If you define relationships, you can define a level once rather than multiple times, which minimizes the sizes of fact tables and their indices. See “Cube-Cube Relationships” in the Advanced DeepSee Modeling Guide.
A special element that is intended for use as the default value of a filter on a dashboard (currently this is their only use). The definition of a runtime variable is an ObjectScript expression that is evaluated at runtime. See “Configuring Settings” in the DeepSee Implementation Guide.
Runtime variables are entirely different from pivot variables.
A measure that enables you to apply a filter that considers the values in the source records. Searchable measures are an InterSystems extension to MDX. In standard MDX, a filter can be based only on members. See “Defining Measures” in Defining DeepSee Models.
A list of multiple MDX items, typically used for rows or columns of a pivot table. The items can be any combination of literal values, members, and tuples. For an introduction, see “Working with Sets” in Using MDX with DeepSee. For reference information, see “Set Expressions” in DeepSee MDX Reference.
A dimension that can be used in more than one cube. That is, more than one cube can use members of the dimension for rows or columns or for filtering. A dimension can be shared formally or informally. If the dimension is shared formally, you can define a compound cube that combines the cubes that use this dimension. See “Defining Shared Dimensions and Compound Cubes” in the Advanced DeepSee Modeling Guide.
The source class is the class that contains the data upon which a cube is based. Every cube has a source class, which is usually a persistent class. A source class has a set of source records. For an introduction, see “Basic Concepts” in Defining DeepSee Models.
See dimension table.
A view of a cube with optional overrides. A subject area uses the fact table and related tables of the associated cube and does not require independent updates. You define subject areas to enable users to focus on smaller sets of data without the need for multiple cubes. See “Defining Subject Areas” in Defining DeepSee Models.
The process of updating the fact table and indices for a cube, based on incremental changes to the source class. See “Compiling and Building Cubes” in Defining DeepSee Models and “Keeping the Cubes Current” in the DeepSee Implementation Guide
See also building a cube.
A simple (but extendable) list of key and value pairs. Term lists provide a way to customize a DeepSee model without programming. See “Defining Term Lists” in the Advanced DeepSee Modeling Guide.
A type of MDX value that consists of an intersection of members. If the tuple refers to each dimension in the cube, the tuple is fully qualified. Otherwise, it is partially qualified.
For an introduction, see “Tuples and Cubes” in Using MDX with DeepSee. For reference information, see “Tuple Expressions” in DeepSee MDX Reference.
Data that is written as text in a human language such as English or French. The iKnow semantic analysis engine analyzes unstructured data. This engine is built into Caché in the same way that DeepSee is. For a general introduction, see “Conceptual Overview,” in Using iKnow.
You can use unstructured data within DeepSee cubes, if the source table for a cube includes a property that contains unstructured data. See “Using Unstructured Data in Cubes (iKnow)” in the Advanced DeepSee Modeling Guide.