Skip to Content

Previously we have talked about the general architecture of SAP BW and the different set of BI tools. However, there are still some tools that were not covered in the previous post and they are a vital part of the BI ecosystem.

SAP Data Services

Enable data integration, data quality, text data processing, data profiling, and metadata management, allowing you to integrate, transform, improve, and deliver accurate, complete data. It provides a trusted information foundation comprising a single development user interface, metadata repository, data connectivity layer, runtime environment, and management console. With it, you can support both operational and analytical data-driven initiatives and access relevant data regardless of type, domain, or source. You can also use it to empower both IT and business users to govern the quality of data.

With this software, you can access, integrate, and process structured and unstructured content from a variety of data sources across your enterprise. These include HANA database, SAP NetWeaver BWapplication. The software lets you:

  • Extract, transform, and load data quickly into a data warehouse to create a complete view
  • Access data from Apache Hadoop to combine unstructured and structured content for new insight
  • Unlock meaning from unstructured documents with native text data processing
  • analyze linguistics to create base metadata, extract information to create semantic metadata, and analyze text data in 31 languages.

Some Quick features about the product:

  • connect to several data sources (DBs, URLs … etc)
  • for each data source there is a set of configurations to be filled
  • in the Data store there are set of objects, meta data, hierarchies, tables … etc
  • you define data flows, build your extraction logic, define sql like logic
  • data services have templates defined, it helps in creating new tables in data sources ( i.e read x table from y system and load into HANA DB )
  • provide data cleansing services in the data quality … set of address cleansing  (country specific and global)
  • there is option to specify the mapping or it will be done automatically ( country and country … etc)
  • they also have a geocoder that converts addresses into lang and lat
  • there is also user defined data cleansing ..  addresses, units … etc
  • text data processing -> entity extraction base transform, Sentiment Analysis
  • you can as well consume web services

Thoughts

  • Integration with Linked data ?!
  • Their suggestion for mappings can be improved, moreover; there are no details about how they map the entities ?! why not to include a semantic mapper for the data ?!
  • The mapping tool again is not user friendly with lots of results, it is a tree based approach with a target and source and mappings are lines drawn between the columns

Business Objects Thing Finder

is a powerful technology for enabling  customized extraction applications. Business Objects ThingFinder analyzes  text and automatically identifies and extracts more than 35 key entity types  out of the box, including people, dates, places, companies or other things  from any text data source, in multiple languages. The ThingFinder  Professional module extends the power of extraction by enabling the  detection and extraction of activities, events and relationships between  entities and giving users a competitive edge with relevant information for their  business needs.I am not sure if they have been merged into one product or as always there are more than product that do the same thing !!

DaSL

DaSL can be firstly understood as high-level query and calculation language for complex analytics, which it is. However, it is also a powerful schema description and transformation language. DaSL allows for specifying complex schema transforms using a high-level business vocabulary rather than specialized and technical relational operators. Just as an SQL query is based on a database catalog, or an MDX query is based on a Cube schema, a DaSL expression is based on a business model exposed by a BI environment, a data source or an application. Instead of exposing physical concepts like tables and columns, a DaSL business model mostly exposes business concepts that are familiar to a BI specialist, to an application developer, or to an expert domain user. Such concepts may include, for instance, a dimension of Customers, properties like Address attached to Customers, a hierarchical dimension of Time with levels Day, Month, Quarter and Year, a measure or key figure of Sales along with its Currency information, etc.

  • Declarative-no «hidden» semantics; expressions are self-sufficient.
  • Transparent –doesnot make irreversible semantic choices on you rbehalf
  • An ambiguous expression doe snot compile
  • Flexible –can adapt to a broad variety of schemas without imposing a specific structure
  • Expressive –can express complex calculations and data topologies
  • Leverages BI culture –flat dimensions, measures, hierarchical dimensions etc. will behave as expected

To Sum up:

  • DaSL is an expression language
  • It aims at proposing a standard language to express business questions in a way that’s completely informal and the semantics of business question can be provided precisely using declarative query language
  • It was created as expressing calculations over the semantic layer was not unified
  • empower the business users and removing IT bottlenecks
  • pushes calculations to HANA and other sources to stop stacking calculation layers
  • extend any universe with a domain specific language for rich analytically computing
  • schema agnostic, doesn’t require a cube, but simple data schema model
  • SQL computes tables, MDX computes star schemas, DaSL computes arbitrary 3NF schemas
  • extract a daSL mdel based on an instance of data from the universe (called data schema), contains all the meta data needed
  • data foundation is the logical view of the database, set of fact tables, a cube coming from OLAP
  • historically in traditional BO stack and OLAP tools we cant have shared dimension, so we have to alias the table and cut the joint .. so to have two tables -> inDaSL we don’t do that as we can specify in what way we want to use the dimension
  • does type inference to make you concentrate on making business queries, if there was any ambiguity you can precise what you want

Search 360 (WIP)

Search 360, a simple to use search service that, unlike today’s products, provides users with meaningful answers to business intelligence questions. Search engines are tools that help users to find documents. But in many cases, the user is not interested in a document, but in an answer to a business question. For example, an answer to the question “How did the revenue per quarter last year relate to the marketing budget of my shops?” might be an ad-hoc generated bar chart together with a data table, only taking into account shops that are actually managed by the user.Search 360 provides capabilities to analyze user queries and translate them into structured queries to arbitrary backend systems, apply some post-processing (e.g., generating a chart) and federate the resulting answers.Search 360 should be integrated into the Business Web platform, be capable of integrating multiple underlying search engines and be consumable from different Business Web applications. It should support the whole process from query entry, search execution to answer display. Moreover, it should serve as a platform to which different groups in SAP Research can contribute their innovative previous and future work on search as components or add new search engines. As such, it will form the basis for a search community in SAP Research to improve collaboration and alignment of groups working on search across different locations. The focus lies at first on assembling existing approaches rather than developing new approaches, surveying the state of the art outside SAP, working together with customers or transferring to existing products.

VISART (WIP)

The VISART project provides a one stop shop for streaming data analytical solutions – an elastic cloud for high velocity BIG DATA analytics.  The ultimate goal of VISART is to hide the complexity of performing BI tasks on top of high velocity BIG DATA. VISART provides easy to consume solutions to specific problems involving streaming data, such as: Production Performance Indicators monitoring or Finacial Instrument Analysis.The goal of the Elastic Complex Event Processing (CEP) is to provide a scalable system for processing of large quantities of information with minimal latency. Elastic CEP allows to scale the processing infrastructure up and down with the amount of data and users using the system at any given moment. Our research objectives are focused on query optimization, operator placement and cost as well as energy driven optimization.Real-Time Visualization project focuses on technical aspects of visual consumption of streaming data and visual operational BI.

ahead – Business Prediction at your Fingertips (WIP)

ahead supports business users to make the right (strategic) decisions by analyzing future scenarios fast and in an intuitive way on a mobile device. With ahead, we want to open predictive analytics to the masses. One major goal of the project is to simplify both the user interaction and the result representation. ahead relieves the user from the need of knowing mathematical or statistical details and from the need of using cryptic languages to do predictive analysis. The quantitative visualization of uncertainty gives instant feedback on the quality of the prediction.As the simplification approach initially limits the power of the application, ahead gradually frees the full power of the algorithms behind according to the user’s experience. By tracking the usage and categorizing levels of experience ahead conveniently adds new features.

SAP Social Media Analytics by NetBase

SAP Social Media Analytics by NetBase is an on-demand, subscription-based solution that can gauge net sentiment – the net result of analytics related to any topic mentioned on social media sites. The application’s sophisticated, natural-language processing engine also extracts insights from postings of social media users.

  • Discover insights and uncover trends in consumer preferences
  • Quantify market perceptions about products, services, and companies
  • Track the success of the social media component of marketing campaigns
  • Extract insights from users’ postings including likes / dislikes, behaviors, and intensity of buyers’ emotions
  • Drill down to specific mentions from the overall sentiment summary chart
  • Create and publish scorecards for end user consumption
To report this post you need to login first.

Be the first to leave a comment

You must be Logged on to comment or reply to a post.

Leave a Reply