by Jelani Harper
The increasing adoption rates of Big Data and data-driven processes are causing the reliance on Business Intelligence (BI) and analytics software to grow. The escalating deployment of these tools is partly responsible for the movement towards the simplification of BI, which is increasingly shifting this technology from the backrooms of IT departments to the front offices of business professionals.
Aiding in the simplification process are the still fledgling mobile and Cloud-based BI options, as well as more popular discovery and self-service tools such as dashboards and visualization software.
This article examines some of the more widespread Business Intelligence software and considers the technical processes they enable to derive informed action from data.
Relational Databases
Relational databases have been one of the most ubiquitous BI tools for several years; a number of recent improvements have made them integral components of BI platforms. In addition to providing a native SQL (Structured Query Language) environment in which to access and store data, improvements in column-store databases enables data to be stored in vertical tables which drastically increase response time, work load capacity, and storage compression.
In-memory technologies have also made relational databases able to store greater amounts of data with swifter access. In-memory databases either forgo or augment conventional disk storage by caching data in RAM or DRAM, significantly boosting storage capacity and data accessing speed. These technologies enable organizations to use relational databases as analytics engines for everything from Big Data to traditional data marts, and facilitate Online Transaction Processing (OLTP) which allows for expedient transactional data. In-memory analytics provide rapid querying speed.
Some of the myriad advanced relational databases include offerings from:
Microsoft: Microsoft SQL Server 2014was made available in mid-October and features a number of new features such as enhanced in-memory OLTP, I/O Resource Governance for management of I/O across multiple databases and workloads, and high availability for in-memory OLTP Databases.
SAP: SAP Hana combines a column store approach with in-memory technologies, and is deployable in Cloud, on-premise, and hybrid versions. It is useful for real-time analytics and transactional data, and facilitates predictive and text analytics.
IBM: IBM has a plethora of SQL-based warehousing options for analytics including DB2, Informix, and Infosphere. Other products include offerings with Neteeza technology such as its IBM PureData Systems, IBM Smart Analytics System, and warehousing solutions on System Z.
Teradata: Teradata has several data warehouses, appliances, and data marts that are specifically designed for Hadoop, analytics, Cloud applications, and other workload types. It also has a range of discovery tools and SQL-based software with which to store and explore Big Data.
Oracle: Oracle Exadata Machine X3-8 scales to petabytes and is ideal for OLTP, warehousing, or a combination of both. Its bandwidth and storage capacity can be increased with the Oracle Exadata Storage Expansion Rack X3-2, which works with other Exadata products as well.
Predictive Analytics
Predictive analytics are the antithesis of conventional BI tools, which are focused on reports of historical data. Predictive analytics use data mining technologies based on the 80/20 rule for algorithms to deliver rapid analysis of future business trends. These technologies are largely used in multi-component platforms alongside with relational databases, dashboards, and interactive visualization tools. The ease of use and celeritous time to insight is ideal for business users. Common use cases for predictive analytics include determining future customer behavior, which makes this software essential for those in sales, marketing, and upper level management.
Interactive Dashboards
Dashboards have played a significant role in the trend towards the simplification of BI, which is also impacted by mobile technologies, self-service and discovery tools, the Cloud, and interaction visualizations. Dashboards have reduced the reporting process from several lengthy pages to a quick glance which details all sorts of information – presented in a variety of graphically effective, unconventional ways – such as metrics and key performance indicators.
Dashboards are only as effective as the data sources they are configured to (which is why they are frequently included in multi-component BI platforms) and may require substantial IT influence to attune them to a particular business unit’s needs. Most discovery tools incorporate dashboards, which also frequently appear in mobile and Cloud-based applications as well. By offering a variety of detailed information pertaining to a specific objective, dashboards play a considerable role in facilitating the movement towards self-service BI and the overall simplification of this technology.
Visualization
Visualization tools extend the visual attraction of dashboards to include several variations on the presentation of data, which Gartner’s Kurt Schlegel states have traditionally included “pie, bar and line charts, including heat and tree maps, geographical maps, scatter plots and other special-purpose visuals.” With the help of IT, end users have substantial autonomy in visually representing data through manipulations involving color, chronology, movement, size, and many others. Users can configure visualization tools to work with existing BI tools or multi-component platforms. This software is valuable for discerning relationships between data, which is why visualization technologies are included in discovery software.
Some of the several BI platforms with dashboards and interactive visualizations include:
SAP: SAP Lumira, available in desktop and Cloud versions, is the principle visualization tool for Sap BusinessObjects. Lumira reduces the need for code while providing intuitive visualizations that enhance agility and complement SAP BusinessObjects Dashboards.
SAS: SAS Visual Analytics utilizes in-memory capabilities with automated analytics to provide ad hoc data discovery for the layman. Visualization features include autocharting that automatically selects the best graph to display data, forecasting, variable relationship explanations, and others.
Tableau: Tableau’s range of products (Tableau Server, Tableau Desktop, Tableau Online and Tableau Public) provide direct connections to databases to enable expedient data analysis with a wide range of dashboards and visual representations that can be published in web browsers and mobile devices. There are also hosted, Cloud-based options.
IBM Cognos: IBM Cognos has a suite of products designed for specific enterprise size and a variety of purposes including BI, analytics, financial performance, and more. Its advanced visualization tools can aggregate a variety of data sources through custom visualizations facilitated by IBM Many Eyes, a community of web-based visualization professionals.
Text Analytics
Although current adoption rates of text analytics are relatively low (no more than 1/5th of BI users deploy this technology), increasing reliance on Big Data could significantly alter this fact. There are many facets of text analytics that are ideal for gauging sentiment data on popular web sites, analyzing fraud-related and security concerns, as well as investigating and classifying data. Various text analytics tools use different technologies including natural language processing, linguistic modeling, and ontologies and taxonomies. Relatively low adoptions rates are influenced by complications such as slang and multiple (foreign) languages, which can obfuscate data and the textual analytics process. As such, this technology is decidedly less user friendly and requires more IT involvement than self-service or discovery tools.
Vendors with viable text analytics components include:
SAS: SAS Text Analytics was specifically designed to derive meaning from unstructured textual data. This product contains specific components for the management of ontologies, sentiment analysis, text mining, and enterprise content categorization.
Oracle: Oracle Text provides SQL-based text analysis of the web, databases, and specific documents. Features include multiple language support, context-based queries, formatting options for search results, and many others.
IBM: SPSS Text Analytics for Surveys was created to provide analysis of sentiment data in surveys using natural language processing. The software incorporates linguistic technologies and reduces the need for coding by automating the categorizing process.
In Summary
There are many universities and professional organizations offering degrees and certificates in BI and analytics now, yet it still remains a complex field. There is no denying the benefits gained from the simplification of these tools to facilitate ease of use for the business user. Most BI platforms include relational databases with predictive analytics, in-memory technologies, dashboards, and visualization tools. Text analytics along with Cloud and mobile applications are swiftly following suit. As a result, analytics and BI software is increasingly influencing decision making in close to real time, while also aiding the overall impact of data in the world today.