Since my last post at oracledmt.blogspot.com, a lot has happened in the world of analytics inside and outside Oracle. Big data and Big Data Analytics are hot topics now. Also, a number of technologies have flourished in the past few years in this space, to name a few: MapReduce, Hadoop, Mahout, Pig, R, and NoSQL database. As a result I felt a new blog was needed with a different theme and focus.
The oracledmt blog focused exclusively on Oracle technology. Because of all the new developments in Big Data Analytics this new blog will go beyond that. It will discuss and illustrate examples with the Oracle stack as well as other technologies. It will cover Big Data Analytics with a special focus on automation. It will also provide information on methodologies and design patterns for solving analytical problems. The effort with this new blog is to move the focus from experts, far too few, towards strategies that can be used by many developers and analysts. The main goal is to help empower developers to create smarter applications that can process, without failure, massive data sets with large number of attributes and adverse data quality. This is a heroic task indeed.
For those wondering, we have also been very busy in the area of analytics at Oracle since my last post. At the technology level we have introduced technologies such as: Endeca, Exalytics, Exadata, Big Data Appliance, In-DB Hadoop, Oracle R Enterprise, and Spatial and Graph analytics. More directly related to my group we have a number of developments:
- The Oracle Data Mining (ODM) database option has been augmented with the addition of Oracle R Enterprise (ORE) capabilities and renamed Oracle Advanced Analytics (OAA).
- We have introduced Dynamic Scoring (a.k.a. Predictive Queries), a database feature that, executing an analytic clause, automates data mining and hides the complexities of the data mining process from the user.
- We have added new in-database algorithms: Expectation Maximization clustering, Singular Value Decomposition (SVD), Principal Component Analysis (PCA), and advanced feature selection and generation for logistic and linear regression.
- Oracle Data Miner (ODMR) is now part of SQL Developer and provides a really nice workflow based data mining workbench which allow mixing SQL and R functionality and increases the productive of data scientists.
Over these years I have also spent a great deal of time working on automating and increasing the ease of use of analytics and big data analytics. This is reflected in different aspects of Oracle Data Miner as well as server-side features like Dynamic Scoring. These ideas and features have also been applied in helping a number of Oracle applications to incorporate advanced analytics, for example:
- Oracle Fusion CRM Sales Prediction Engine
- Oracle Fusion HCM Workforce Predictions
- Oracle Spend Classification
- Oracle Adaptive Access Manager
- Oracle Airline Data Model
- Oracle Communications Data Model
- Oracle Retail Data Model
- Oracle Security Governor for Healthcare
- Oracle Communications Social Network Analytics.
I have migrated the content from the oracledmt blog to this new one. However, fixing some details in the posts is still a work in progress. I hope you enjoy the new look and content!