By Charlie Berger, Advanced Analytics-Oracle on Jul 26, 2015
Big Data Analytics with Oracle Advanced Analytics:
Making Big Data and Analytics Simple
Oracle White Paper | July 2014
Executive Summary: Big Data Analytics with Oracle Advanced Analytics
The era of “big data” and the “cloud” are driving companies to change. Just to keep pace, they must learn new skills and implement new practices that leverage those new data sources and technologies. Increasing customer expectations from sharing their digital exhaust with corporations in exchange for improved customer interactions and greater perceived value are pushing companies forward. Big data and analytics offer the promise to satisfy these new requirements. Cloud, competition, big data analytics and next-generation “predictive” applications are driving companies towards achieving new goals of delivering improved “actionable insights” and better outcomes. Traditional BI & Analytics approaches don’t deliver these detailed predictive insights and simply can’t satisfy the emerging customer expectations in this new world order created by big data and the cloud.
Unfortunately, with big data, as the data grows and expands in the three V’s; velocity, volume and variety (data types), new problems emerge. Data volumes grow and data becomes unmanageable and immovable. Scalability, security, and information latency become new issues. Dealing with unstructured data, sensor data and spatial data all introduce new data type complexities.
The world of traditional advanced analytics is inherently wrought with weak points of server analytical servers, data movement, data duplication, data security, different tools for different analysis purposes and ever growing information latency. Significant challenges become especially clear during predictive models deployment projects when attempts to mash together the separate worlds of DBAs, IT and the database with the worlds of data scientists, advanced analytics and separate analytical servers highlight these differences.
Traditional data analysis typically start with a representative sample or subset of the data that is exported to separate analytical servers (SAS, SPSS or R, etc.) that have been especially designed for statisticians and data scientists who analyze the data. These analytics range from simple descriptive statistical analysis to advanced, predictive and prescriptive analytics. If a data scientist builds a predictive model that is determined to be useful and valuable, then IT needs to be involved to figure out deployment and enterprise deployment and application integration issues become the next big challenge. The predictive model(s)—and all its associated data preparation and transformation steps—have to be translated to SQL somehow and recreated inside the database in order to apply the models to make predictions on the larger datasets maintained inside the data warehouse. Or, sometimes vendors will change for additional model deployment products. This model translation phase introduces tedious, time consuming and expensive manual coding steps from the original statistical language (SAS, R, and Python) into SQL. DBAs and IT must somehow “productionize” these separate statistical models inside the database and/or data warehouse for distribution throughout the enterprise. This is where many advanced analytics projects fail. Add Hadoop, sensor data, tweets, and expanding big data reservoirs and the entire “data to actionable insights” process becomes more challenging.
Not with Oracle. Oracle delivers a big data and analytics platform that eliminates the traditional extract, move, load, analyze, export, move load paradigm. With Oracle Database 12c and the Oracle Advanced Analytics Option, big data management and big data analytics are designed into the data management platform from the beginning. Oracle’s multiple decades of R&D investment in developing the industry’s leading data management platform, Oracle SQL, Big Data SQL, Oracle Exadata, Oracle Big Data Appliance and integration with open source R are seamlessly combined and integrated into a single platform—the Oracle Database.
Oracle’s vision is a big data and analytic platform for the era of big data and cloud to:
· Make big data and analytics simple (for any data size, on any computer infrastructure and any variety of data, in any combination) and
· Make big data and analytics deployment simple (as a service, as a platform, as an application)
Oracle Advanced Analytics, an option to the Oracle Database Enterprise Edition 12c, extends the database into an enterprise-wide analytical platform for data-driven problems such as churn prediction, customer segmentation, fraud and anomaly detection, identifying cross-sell and up-sell opportunities, market basket analysis, and text mining and sentiment analysis. Oracle Advanced Analytics empowers data analyst, data scientists and business analysts to more extract knowledge, discover new insights and make informed predictions—working directly with large data volumes in the Oracle Database.
Oracle Advanced Analytics offers a wide library of powerful in-database algorithms and integration with open source R that together can solve a wide variety of business problems. Oracle Advanced Analytics in-database analytical functionality and predictive algorithms are all accessible via the SQL API data mining and statistical functions (Oracle Data Mining), via the SQL Developer 4.1 Oracle Data Miner “workflow” GUI extension, and via integration with the R open source statistical programming language (Oracle R Enterprise). Oracle Advanced Analytics’ SQL transparently layer pushes down R functions to equivalent ODM and SQL functions for in-database performance and scalability. By integrating big data management and big data analytics into the same Oracle Database 12c platform, Oracle eliminates data movement, reduces total cost of ownership and delivers the fastest way to deliver enterprise-wide predictive analytics solutions and applications.