Birst Review Summary
Birst is a cloud (public or private) based BI platform that addresses the needs of both production oriented BI and end-user data visualization and querying. The company calls this a 2-Tier approach, and it is possible because of the excellent data management architecture of the platform.
Many contemporary data visualization platforms are, in effect, point solutions. They confuse the data picture rather than clarify it, by creating their own versions of the truth within a database of their own. They also do not cater for production BI needs which, although currently somewhat unfashionable, are nonetheless here to stay. Production reporting (weekly sales reports for example), executive dashboards and other ‘fixed’ information needs will always be with us, despite the current fascination with all things visual.
So Birst competes very well with the ease of use of products such as Tableau and Qlik Sense, while also providing a true enterprise BI facility, with support for regular, production oriented reporting.
The impressive feature list of Birst includes native and offline mobile support, embedded predictive tools, intelligent data functions, automated data refinement, drag and drop data visualizations, and quite unique levels of data management. This integrated environment means that everyone gets the same version of the truth, that predictive models can be embedded into dashboards and charts, and that there should never be data ambiguity.
The architecture of Birst is sophisticated and involves several layers, most of which concern themselves with the data. While we all want to create attractive visualizations, it is the underlying data architecture that empowers accurate and meaningful data visualization.
The total cost of ownership of Birst is reported by users and analyst firms to be low, and of course the time to value is very low. Birst is a contemporary alternative to most conventional enterprise BI platforms, while delivering modern user interfaces and well governed self-service capability.
User Interface
Birst provides an easy-to-use interface for data exploration, discovery and visualization. It supports the creation of charts, dashboards and high definition reports. The self-service capability of Birst equals many of the products which specifically address this need. An Open Client interface is also provided if users want to use other tools, such as Excel or Tableau.
The visual discovery interface is mainly navigated using drag-and-drop, and a recommendation engine will suggest the most appropriate visualizations for the data. It supports filtering, user created metrics, instant metrics, and an intelligent search capability. Available visualization formats include: column, bar, line, spline, area, area spline, points, scatter, bubble, pie, funnel, pyramid, list tables, crosstabs, and geo maps. The Expression Editor provides the ability to create BQL (Birst Query Language)-based report expressions that can be used to create more insightful visualizations, and users can also choose to limit data to “Top N” data points.
Dashboards are created in a WYSIWYG, drag-and-drop environment, and Birst dashboards and widgets are rendered in HTML5, so they automatically resize for a responsive, optimized experience wherever they are used, on a laptop, desktop, or tablet.
Birst also includes a report designer for advanced pixel-perfect report creation, enabling highly formatted report creation typically used in production-delivered reports. Examples of rich formatting include: conditional formatting, conditional display, duplicate suppression, and null value replacement. Embedded images and sub-reports in various bands are supported. Reports are compiled into Java byte code for fast and direct execution. No interpretation at runtime is required, and server-side report caching enhances performance.
The Birst platform includes a predictive analytics engine. In contrast to conventional data-mining environments, data does not have to be moved; instead datasets for model training and scoring are generated directly from the Birst Unified Business Model.
Birst’s advanced analytics capabilities leverage the ROLAP engine for data preparation. The modeling engine makes full use of aggregates and derived measures. Sophisticated new measures are defined and calculated on the fly as inputs into the modeling process. Share, time-series and dimensional breakout metrics are used to enrich the information. The use of OLAP-style measures for modeling enables the addition of complex and highly predictive behavioral calculations. For each modeling task, Birst automatically evaluates a comprehensive set of algorithms. Supported algorithms include linear and logistic regression, decision trees, feed-forward neural networks, support vector machines and rules/regression trees. Modeling scores are written directly back to the User-ready Data Store, ready to be used in ad hoc queries and dashboards or to be fed into additional processing (for example, list generation). Both rules-based and model-based recommendations can be combined into complex decisions. Birst also delivers tight integration with the R statistics package, making it easy to deploy R-based measures to any number of users. Birst measures can make calls to the R server, submit data for processing, and retrieve the results to present to users. By leveraging the integration with R, users can greatly augment the already robust advanced analytics capabilities available in Birst out of the box.
Data Management
It is the data management and manipulation capabilities that really distinguish Birst. It provides data extraction and connectivity options for a wide variety of databases, flat and structured files, analytic databases, and popular cloud and on-premises applications.
Birst supports the extraction of entire database tables or views, and the extraction of subsets of data using custom SQL queries. Birst Connect, a Java application sitting on-premises, can be used for both bulk data extraction and for connecting in real-time (see Live Access below.) Birst extraction tasks can be scheduled either using a built-in scheduler or an external OS scheduler. In addition to extraction of data from all relational and analytic databases (SQL Server, HP Vertica, Teradata, Amazon Redshift, etc.) and applications, Birst supports uploading delimited flat files, Microsoft Excel and Access database files. Structured data is extracted and uploaded in a tabular format of columns and rows per data sources. These extract and live query capabilities also extend to modern (unstructured) and big data sources such as Hadoop Hive, Cloudera Impala and Cassandra. In all cases, data is transferred securely to Birst using secure authentication and compression techniques.
For even faster deployments and zero connector maintenance, Birst offers a set of pre-built connectors to over 100 popular cloud and on-premises business applications like Salesforce, SAP, Eloqua, and others. These connectors (leveraging web-service APIs and/or JDBC) have been designed to extract standard and custom objects, or even specific columns, from the respective applications and are maintained over time to ensure connectors stay up to date with application changes.
Birst also offers Live Access (or real-time query) capabilities to directly query on- premises data sources, like XMLA cubes, existing enterprise data warehouses, data marts, applications, and data lakes. On-premises data stores or applications are queried in real-time without the need to first extract and load the data into the cloud. This helps to bridge the gap between centralized and decentralized teams, enabling enterprises to leverage their existing investments in data warehouses, data marts and XMLA cubes (their centralized BI assets) while still leveraging Birst for combining decentralized data and creating the essential Business-ready Data Tier. Live Access connects to on- premise data sources directly, in real-time, and transfers query results securely via the Transport Layer Security (TLS).
Birst’s cloud analytics engine delivers automated data integration capabilities for standard data integration needs, and a developer-friendly scripting (ETL) language for more complex needs. An example of this automated refinement is Birst’s automated time-series measures, where all measures are automatically available by common time-series dimensions, like trailing 12 months, trailing 3 months, etc. Since Birst is a single platform, all data integration routines are developed, tested, and enabled in production from a single web browser, without any work in a different application or desktop-based tool.
Birst’s User-ready Data Store seamlessly combines different sources of data. It is designed and optimized for ROLAP-style analytics, providing a Kimball-style star schema with a multidimensional view of all data. In addition, Birst supports Type 1 and 2 slowly changing dimensions, conformed dimensions, and manages snapshots and time-based transformations automatically. Data loading and updates are done through incremental processes with built-in change detection.
Embedding
Birst supports powerful embedding options and empowers software providers to quickly and seamlessly embed business analytics into their applications and leverage Birst to differentiate from their competitors, deliver more value to their customers, and create new revenue streams. Birst web services enable programmatic administration of a Birst solution and tight integration into other applications or portals.
Birst works within the larger enterprise applications ecosystem and provides the ability to embed reports and dashboards in the cloud or other applications. The SSO framework supports session parameters to dynamically control access privileges and data visibility to those logged in to the application. For authentication, Birst also supports OpenID and SAML 2.0.
Birst’s web services APIs extend Birst as an open platform for embedding into any SaaS or web application. Birst supports all methods of web services to receive data whether it’s REST or SOAP. The outbound Birst web services API is SOAP- based and can work with any programming language that supports web services. Web services range from calls to managed users and metadata to services for running queries.
Competition
Birst successfully bridges the need for enterprise production business intelligence, and self-service data visualization and exploration. This makes it fairly unique, and so it is difficult to compare other products. However here are several products that might be compared with Birst:
- Tableau is purely a self-service data visualization platform, and while it might be used to augment Birst (unnecessary in our view), it in no way compares with the depth and breadth of Birst.
- Qlik Sense also focuses on data visualization and discovery, but comes with excellent extensibility, and a unique associative data engine. It is probably the most sophisticated in this genre of BI platforms. However Qlik Sense does not support production reporting.
- DOMO is targeted at executive and management information needs and is weaker than Birst in terms of the data exploration and discovery capability, and the underlying data architecture.
- GoodData is perhaps the nearest match to Birst, with good data visualization tools, a solid data architecture, and some level of support for advanced analytics (R integration).