Abstract: A system and method include input of data records to a first trained predictive model to obtain a predicted value associated with each input data record. A model region is then associated with each of the input data records based on the first trained predictive model, the input data records and the predicted values. Enhanced input data records are generated by, for each model region, adding derived values of engineered features associated with the model region to input data records associated with the model region and default values of the engineered features associated with the model region to input training records not associated with the model region. The enhanced input data records are input to a second trained predictive model to obtain an enhanced predicted value associated with each input data record.

Abstract: A method for sharing access to data from a cloud-based analytics engine may include receiving, at a first client, a message from a second client. The message may include an image of a data presentation and metadata associated with the data presentation. The first client may respond to a selection of the image of the data presentation by accessing the cloud-based analytics engine to retrieve, based on the metadata associated with the data presentation, data associated with the data presentation. The data presentation may generated for display at the first client based on the data retrieved from the cloud-based analytics engine. The data presentation displayed at the first client may be interactive to enable an application of one or more modifications to the data presentation displayed at the first client. Related systems and articles of manufacture are also provided.

Abstract: A method for accessing data stored in a database may include receiving, at an application, a request to load a data story that includes a software widget configured to create, based on a portion of data stored in a database coupled with a cloud-based analytics engine, a data presentation providing a visual representation of the portion of data. In response to the request to load the data story, a scripting engine associated with the application may execute a first programming code associated with the software widget. The first programming code may be executed to retrieve, from the cloud-based analytics engine, visualization data for rendering the data presentation. An in-app web browser associated with the application may execute a second programming code associated with the software widget. The second programming code may be executed to render, based on the visualization data retrieved from the cloud-based analytics engine, the data presentation.

Abstract: A system includes reception of a first query from a first user, identification, in response to reception of the first query, of a first plurality of data entries, each of the first plurality of data entries associating the first user, the first query and a respective visualization type with a respective counter value, determination of one of the first plurality of data entries associated with a greatest respective counter value of the counter values of the first plurality of data entries, determination of a respective visualization type of the one of the first plurality of data entries, and presentation of a visualization of the determined visualization type of a first result set corresponding to the first query.

Abstract: According to some embodiments, a business intelligence data store contains semantic layer information. A business intelligence language platform may receive a business intelligence language query expression from a user. The business intelligence language platform may then access the mapping information in the business intelligence model data store to compute an output type describing a result structure entity for the query expression. According to some embodiments, the output type may be used to automatically recommend at least one visualization chart to the user. Moreover, in some embodiments the output type is annotated with meta-data information about a business object.

Abstract: A system includes acquisition of data from a first data source, identification of logical entities of a first data model of the first data source based on the data of the first data source, identification, for one or more of the logical entities of the first data model, of a respective knowledge vault entity described by first metadata, and generation of second metadata describing a plurality of semantic layer entities, the second metadata describing a semantic layer entity comprising at least one semantic attribute of a knowledge vault entity corresponding to the semantic layer entity.

Abstract: According to some embodiments, an abstraction engine may determine a data manipulation definition generated by a creator platform to accesses a database via a first connection protocol (e.g., ODBC). The abstraction engine may then automatically analyze the data manipulation definition to discover a connectivity parameter (e.g., a DSN, a DBMS type, a DBMS host name, a port, etc.) associated with the access to the database via the first connection protocol. The data manipulation definition may then be stored along with the connectivity parameter as a meta-connection into a metadata repository. A consuming platform may retrieve the meta-connection from the metadata repository translate the meta-connection into the database manipulation definition to accesses the database via a second connection protocol (e.g., JDBC).

Abstract: A system includes presentation of a visualization of a result set, the visualization including a plurality of graphical representations, each of the plurality of graphical representations associated with a respective data point comprising at least one dimension value and at least one measure of the result set, reception of a selection of a first data point of the plurality of data points, presentation of an interface to receive a one or more comments in response to selection of the first data point, reception of a first one or more comments, and storage, in response to reception of the first one or more comments, of an identifier of the first data point in association with the first one or more comments.

Abstract: A method for configuring an algorithm for processing data in a database may include generating a plurality of rules by converting a concrete syntax representation of one or more constraints associated with a parameter of the algorithm into an abstract syntax representation of the constraints. A first input value associated with the first parameter may be received from a client via a user interface at the client. The user interface may be updated based on the first input value. The user interface may be updated by applying a first rule from the plurality of rules. The update to the user interface may include a change to a visibility, a modifiability, and/or an allowed value of a second parameter of the algorithm. Related systems and articles of manufacture, including computer program products, are also provided.

Abstract: Some embodiments provide a non-transitory machine-readable medium that stores a program executable by at least one processing unit of a device. The program receives data associated with a visual presentation that includes several visual elements. The program also identifies a first set of visual elements in the several visual elements having a first type and a second set of visual elements in the several visual elements having a second type. The program further renders the first set of visual elements in a first layer of the visual presentation using a first rendering engine. The program also renders the second set of visual elements in a second layer of the visual presentation using a second rendering engine.

Abstract: A computer-implemented method for disaggregating data in a spreadsheet or other data structure hosted on a client computing device includes providing an analyzer module as a plug-in or add-in to a spreadsheet application on a client computing device. The analyzer module locally performs disaggregation and aggregation processes on data in the spreadsheet on the client computing device. The method includes parsing, by the analyzer module, content entered in a subject cell of the spreadsheet to determine if sets of special syntax characters corresponding to a disaggregation instruction or command precede an amount in the entered content, and accordingly disaggregating or aggregating data in the spreadsheet.

Abstract: According to some embodiments, a business intelligence language expansion and compilation platform may receive a business intelligence language expression, containing at least one macro sub-expression from a user. Information may then be accessed in a business and data models data store to perform expansion and compilation operations on the received business intelligence language expression. As a result of these operations, the system may provide a native query to a data source database. A business intelligence language result set manager may retrieve data from a data source database responsive to the native query and output a business intelligence language result set to the user in response to the business intelligence language expression.

Abstract: A system includes reception of name data, determination, for each of a plurality of name properties, of an associated property value based on the name data, determination of a gender classification based on the property values, and, for each property value, generation of a substitute property value based on the property associated with the property value and the gender classification.

Abstract: Provided are a system and method for semantic generation and enrichment. In one example, the method includes displaying a plurality of tiles on a user interface, each displayed tile visually representing a data entity obtained from a data file, detecting, via the user interface, a selection of a first displayed tile visually representing a first data entity and a second displayed tile visually representing a second data entity, and in response to detecting the selection via the user interface, generating a new data entity based on the first data entity visually represented by the first displayed tile and the second data entity visually represented by the second tile, and displaying a new tile visually representing the newly generated data entity. Accordingly, an end user can provide semantic creation and enrichment to visually displayed data without the need for an administrator.

Abstract: Some embodiments provide a non-transitory machine-readable medium that stores a program executable by at least one processing unit of a first device. The program establishes a collaboration session with a second device. The program also receives, from the second device, analytics data shared through the collaboration session and contextual data associated with the analytics data. The program further provides additional data for display on the first device based on the contextual data associated with the analytics data.

Abstract: Embodiments relate to techniques for performing data blending operations across multiple different data sets comprising data structures with columns and rows. The data sets may be classified and displayed in a visualization (i.e., chart) in a client interface. Columns and rows from the blended data sets may be mapped together (i.e., linked). Updates to the visualization, including adding elements from the data sets, may trigger a data blending process on the backend server in communication with a database. The server may blend the specified data by generating a runtime artifact representing a calculation graph for the blend operation and query the database to retrieve a resulting data set. The data blending operation may comprise collapsing dimensions of a primary data set with linked dimensions of a secondary data sets into a blended column and aggregating values of measures in rows of the blended column of the resulting data structure.

Abstract: Embodiments relate to performing data blending operations across multiple different data sets comprising data structures with columns and rows. Columns of data sets to be blended may be linked together. Filters may be applied to data sets before the data blend operation is performed to specify which columns to be displayed in a visualization at a client interface. A direct filter may be applied to one of the data sets to obtain a filtered resulting data set. Data elements of the filtered resulting data set can be identified that correspond to the linked columns of the data sets to be blended. The results of applying the direct filter may then be used as the filtering criteria for an indirect filter to filter a second data set. The results of applying the direct and indirect filters may then be combined together in the data blending operation.

Abstract: Some embodiments provide a non-transitory machine-readable medium that stores a program executable by at least one processing unit of a first device. The program generates, at the first device, a snapshot of a digital content page being presented by a second device. The program also provides a set of annotation tools for annotating the snapshot of the digital content page. The program further adds a set of annotations to the snapshot of the digital content page using the set of annotation tools.

Abstract: In one embodiment, a method receives a query for data in a database system and calls a plurality of engines to analyze information for the query. A calculation graph is generated from at least a portion of the plurality of engines where each of the at least a portion of the plurality of engines add a node to the calculation graph based on analyzing of the information for the query. Then, the method executes the query by calling the nodes of the calculation graph. Each node uses metadata added to the node from a respective calculation engine to perform a calculation for the node. Then, a result of the query is output based on the respective calculations performed by the nodes.

Abstract: Embodiments relate to transforming data views for analysis and/or collaboration purposes. An original data view is displayed on a document page in response to a database query. A filter operation is performed by selecting a specific subset of that original data view. A copy of that transformed data view portion including a tuple, is displayed in a separate analysis space (e.g., on the same document page, on another page, or within a collaboration panel). In one example, a user clicks/drags a subset of a table data view for transformation and copying. When the selected and transformed data view portion is copied to the analysis space, relevant associated context information (e.g., the underlying query, measures, dimensions, chart legends, formatting) may be retained for later access. To conserve the limited space typically available to a collaboration panel, certain embodiments may simplify presentation by reproducing only the characters of the transformed data portion.