Wednesday Jul 06, 2016

Moving Towards Analytic-Driven HR: How Sophisticated do you need to be? How sophisticated are you?

Moving Towards Analytic-Driven HR How Sophisticated do you need to be? How sophisticated are you? Maybe it’s a good time to ask for a bit of self-reflection. In this post, I describe some characteristics of four HR cultures:

1. This HR organization produces operational reports showing basic efficiency and compliance measures. Its focus is on distributing accurate, pre-formatted results in a timely fashion.

2. This HR culture has expanded beyond operational reporting to produce advanced reports that show benchmark results and provide decision support. These reports are often in the form of dashboards supported by multi-dimensional analysis.

3. This HR group does statistical modeling and root-cause analysis to address business problems. Their advanced analytics proactively identify issues and recommend actionable solutions.

4. This HR unit produces predictive analytics using scenario planning, risk analysis and predictive models. These analytics are often integrated with the organization’s strategic planning.

Noticeably, these four HR cultures have been presented in increasing “level of sophistication.” Now, ask yourself these questions: • Where are you - which of the described HR cultures best describes your organization?

• Where would like to be - which of the described HR cultures is most likely to be viewed as a value-added business partner?

• If you are not at the desired level of sophistication, what do you need to do to get there?

I will be posting more thoughts on attaining the desired “level of sophistication.”

Wednesday Jun 29, 2016

Sourcing DV from Essbase data - without OBIEE Server

Need to create some nice DV visualisations on your Essbase cube data, but you don't have a BIEE server connected to Essbase... A simple but robust way to do this is to leverage SmartView inbetween Essbase and Oracle DV.
Create an Excel spreadsheet with your Essbase data using SmartView. Make sure you have selected a few SmartView options to make it consumable by Oralce DV (see video), and simply import that file into Oracle DV.
Check out this short video example - turn sound on.

Friday Jun 03, 2016

What's new in DV Desktop 1.1 (

A new version of Oracle DV Desktop was just released on Jun 3rd 2016. Build 05 28 011143, aka V1.1 of DVD. If you were already using DV Desktop, you should get a prompt to upgrade, if not, just visit the OTN page here and download the software.

Along with bug fixes, few new items came with this release. below is a summary list, we will publish more detailed entries/videos soon to illustrate some of these :

  • New beta connections types for more datasources : Greenplum and RigthNow are added as Beta connections.

  • Connectivity to Oracle Dbaas : using the oracle DB connection type, users can directly create direct connections to their DBaaS services (DB as a Service sources).

  • Ability to invoke database specific SQL syntaxes directly from within a calculated item, using EVALUATE function. This enables to fully leverage the underlying power of any database, by function-shipping the calculation process directly down to the database and retrieving only the result rows. For example, clustering tens of millions of transaction IDs into only few clusters : the clustering process can be pushed down in the database, and DVD will only retrieve and visualize the few cluster rows produced by the database calculation. This capability is now available OOB within the 'New Calculation' dialog box, at the time of building analysis, using EVALUATE function. See this video describing the feature

  • The 'Attribute' Function ATTRIBUTE function is available in custom calculations, and turns expressions (either measures or attributes) into attributes. If you are familiar with Answers, and ever used the 'Treat as an Attribute' Flag in the Answers expression editor, ATTRIBUTE function is the syntax behind this functionality. It is useful when computing aggregates on rows that are grouped by values of another aggregate. For example, how many customers have made 1,2,3, N orders ? How much profit in each revenue by customer decile... etc.Syntax : ATTRIBUTE(<expression> BY <attribute list> WHERE <predicate>) A more detailed blog entry will come soon about this feature

Using Forecast Function in Oracle DV / OBIEE 12c

Oracle DV and OBIEE 12c offer a right click interaction to include forecast data in many visualizations.

But it also let users manually configure and edit Forecast functions as custom calculations. This Forecast function is based on a R script ran by Oracle BI Server, and comes with various options and parameters enabled. For example, it allows to define the number of periods to forecast, the type of forecasting model to use (Arima, ETS) and what specific parameters for this model (error type, seasonality, trending, Box Cox, Damping...) as well as what output data to produce (value, confidence levels high and low bounds). All these options combine in a range of various possibilities for calculating the most appropriate forecast information for the use case.

A free pre-built DV Destkop project showing various combinations of these options has been posted on the Oracle Technology Network Dataviz example page here (scroll to project example 'Forecast Syntax Examples') and this YT video gives a very brief introduction of how to use this function :

The FORECAST() Logical SQL Function takes a measure parameter e.g. revenue and a variable list of time dimensions. Optional column aliases can also be used. The default FORECAST XML (filerepo://obiee.TimeSeriesForecast.xml) script file on your server or laptop can be overridden by specifying a new one in the options string. 

Syntax : FORECAST( <measure_expr>, (<time_dimension_expr>), <column_name> , <options>, [<runtime_binded_options>])

- measure_expr represents the measure, e.g. revenue data to forecast.
- time_dimension_expr the time dimension to forecast. One or more columns may be provided.
- column_name represents the output column name for forecast.
- options is a string list of name/value pairs separated by ';'

Option Name

numPeriods the number of periods to forecast Integer

the confidence for the prediction

Integer (1 to 99)


the model to use for forecasting



if TRUE use box cox transformation



the Box-Cox transformation parameter. Ignored if NULL or FALSE.


(ETS model). if TRUE, use damped trend, ie reduce effect of recent trends.


(ETS models) : controls how the nearest prior periods are weighted in the output

additive('A'), multiplicative('M'), automatic('Z')


(ETS models) : controls how the effect of trend is modeled in the output

None('N'), additive('A'), multiplicative('M'), automatic('Z')


(ETS models) : controls how seasonal effects are affecting the model outputs.

None('N'), additive('A'), multiplicative('M'), automatic('Z')

modelParamIC Information criterion to be used in comparing and selecting different models and select the best model. 'ic_auto', 'ic_aicc', (corrected Akaike IC),'ic_bic‘ (Bayesian IC), 'ic_auto'(default)

- runtime_binded_options is an optional comma separated list of runtime binded colums and options.

Some Examples from the DVD Project :

Selecting prediction confidence interval, and showing low end and high end bounds

ETS vs ARIMA : what are the differences ?

Playing with ETS Trending and Seasonality parameters, what does it mean :

Find out more by downloading the example here (scroll to project example 'Forecast Syntax Examples').

Thank you.

Wednesday May 18, 2016

Oracle Excellence Award: Oracle Cloud Platform Innovation - Business Analytics

At Oracle, we know that our customers are our best innovators, greatest teachers, and biggest inspirations. With Oracle technology as the backbone, our customers are deploying revolutionary solutions, establishing and refining best practices, and becoming thought leaders in their industries. The Oracle Excellence Awards celebrate their wins, honor their achievements, and inspire them to achieve even greater success in the future.

This year's Oracle Cloud Platform Innovation – Excellence Awards will honor customers and partners who are creatively using Oracle products. Think you have something unique and innovative with Oracle Business Analytics products?

We'd love to hear from you! Please submit today in the Business Analytics category. This year’s nomination process is a little bit different in the sense that the forms are fully online.

The deadline for the nomination is Monday, June 20, 2016. Win a free pass to Oracle OpenWorld 2016!

Submit Now

For details on the 2015 Big Data, Business Analytics, and Data Integration Winners:, Caixa Bank, Serta Simmons Bedding, Skanska, Scottish & Southern Energy and Tampa International Airport, check out this blog post

Advanced Analytics in Oracle DV

Yes, Oracle DV gives you advanced analytics literally at the click of a button. Advanced analytics operations such as clustering, outlier detection, trend lines and forecast are built-in, and they can be simply dragged dropped on to your charts.

However access to advanced analytics is not limited to these functions exposed in the UI. You can leverage the advanced analytics framework to invoke any custom R function and seamlessly blend its results with rest of your analysis. The EVALUATE_SCRIPT function of Oracle DV can be used to execute an R script registered in the script repository.  Its powerful and easy to use.

For example, this video shows an example of using custom R script to decompose a time series data to its seasonal, trend and remainder components.

Following video is another example where use R to perform market basket analysis (rule mining) on fast foods transactions data.

Take a look at some of these examples downloadable from Oracle DV samples page on OTN.

Tuesday May 17, 2016

Oracle DV : Advanced custom SQLs as source for Oracle Data Viz

Oracle DV lets you define custom data sources as free SQL statements to any database connection (ORCL, SQLServer, TD...). These SQL statements can include any advanced syntaxes specific to the underlying database.

The recorded example below shows defining a custom SQL data-source from an Oracle DB table, and then extending it by adding a dynamic predictive clause in the SQL. Resulting data-set interacts live with user filters and allows to build impact-full visuals while leveraging the full power of underlying highly scalable databases..

Thursday May 12, 2016

Oracle Big Data Discovery Cloud Service is GA now!

Get Value from Big Data Faster than Ever

With the availability of Oracle Big Data Discovery Cloud Service, customers now have immediate access to a Data Lab, offering them the full power of Big Data Discovery in the Oracle Cloud, enabling unprecedented speed and agility for big data analytics.


Oracle Big Data Discovery Cloud Service (BDDCS) offers all the capabilities of Big Data Discovery 1.1, as well as push-button provisioning and easy cloud life cycle management, including:

  • Data Lab in the Cloud: Leading organizations are disrupting their industries by unlocking commercial value in big data. They start with data labs, complete environments for agile big data analytics experimentation and innovation. BDDCS removes the burden of complex in-house architectures and capital expense, making the data lab immediately available so customers can rapidly engage in broad-scale big data innovation. And because BDDCS is highly visual and intuitive, customers can easily extend their big data team to business users and analysts, who can find, explore, transform, and analyze data at scale, then easily share results. This accelerates the pace of innovation, while reducing the need for highly skilled and very scarce resources.
  • Power and Flexibility: As momentum continues to shift to the cloud, customers increasingly expect highly powered, robust environments as the foundation for their solutions. BDDCS leverages the power of Oracle Big Data Cloud Service and Apache Spark to turn raw data into business insight in minutes. Oracle Big Data Cloud Service includes Oracle R Advanced Analytics for Hadoop (ORAAH), Big Data Spatial and Graph, Big Data SQL (option), and more, giving customers the power and security they require, and the flexibility to take their analytics as far as they want.
  • Simple Administration: Along with power and flexibility, BDDCS offers easy, automated lifecycle management, to give customers the best possible set of services for innovation with big data. Customers can provision a new Big Data Discovery Cloud Service instance with the push of a button, monitor availability and usage, easily apply updates to instantly take advantage of the latest product capabilities, schedule automatic backups into Oracle Cloud Storage Service, and more.

Learn more 

Friday May 06, 2016

Oracle Data Visualization Desktop Gallery

Samples and Examples are now available on OTN. Visit the DVD public gallery for pre-built Data Visualization applications

Monday May 02, 2016

Oracle Data Visualization Desktop is available now!

Data Visualization Desktop (DVD) is a standalone capability that provides rich, consumer-style analytics, including intelligent data visualization, self-service data discovery, visual storytelling, and predictive analysis, as well as out-of-the-box access to numerous data sources and built-in data preparation.

 Highlights and benefits include:

  • Easy, Engaging Visual Analytics: Oracle Data Visualization blends drag-and-drop simplicity with powerful analytics, making it fun and easy for anyone to blend, visualize, and gain new insight from their data.  DVD makes visual analytics even more accessible through a quick download and install, allowing anyone to have a local best-in-class visual analytics capability that includes new visualizations, free form layouts, custom coloring and highlighting, and more.  See the new features video.

  • Anytime, Anywhere:  DVD complements DVCS, BICS, and BI 12c, making the same rich experience and technology available for personal use on the desktop.  DVD is ideal for organizations that are still sorting out their cloud strategy, offering easy access to visual analytics that creates a natural bridge between premises and cloud.  Equally, it offers cloud customers a simple solution for local analytics, enabling people to work anywhere they need to.     

  • Push-Button Data Management:  Point-and-click data blending allows users to blend data from a variety of sources – Oracle and other SaaS applications, on-premises systems, external sources and personal files, including:  OTBI/E, Oracle database, Spark, SQL Server, MySQL, DB2, Teradata, Redshift, Hive, Impala, MongoDB, SybaseIQ, and Salesforce.  DVD also offers integrated data preparation capabilities, enabling anyone to quickly clean and enhance data as it is loaded, as well as export to CSV files.

  • Predictive Analytics:  Analytics has progressed from providing oversight to offering insight, and now to enabling foresight.   Oracle Data Visualization supports that progression, delivering embedded predictive capabilities that enable anyone to see trend lines and other visuals with a click, and extend their analysis using a free R download.  

Download DVD from OTN

Monday Apr 25, 2016

What’s new in Big Data Discovery 1.2?

Oracle Big Data Discovery offers a new approach for turning big data into commercial value, fast.  BDD makes exploring big data as easy as shopping online, allowing ordinary analysts to find, improve, analyze, and share data without needing the technical skills of a data scientist. No other big data solution in the industry today covers the complete analytics lifecycle in a single product or makes it as quick and easy to stand up a fully functional data lab.

BDD 1.2 takes on the biggest problem in big data: harnessing the knowledge and skills of business analysts to solve unprecedented problems with new data. This release features major new functionality all across the product, along with hundreds of performance improvements that make the platform faster and more reliable than ever.


· Wrangle + Reshape: Big Data Discovery 1.2 adds support for aggregation and join transformations, allowing users to group and blend data of any size interactively using Spark. Combining open source innovation with Oracle engineering, BDD provides an easy-to-use visual interface that guides users through these essential data shaping tasks—allowing analysts to sculpt new datasets to power new analytics, without the need for IT’s technical skills.

· Curate + Locate: Big data is messy data. Making data usable involves not just cleaning the data itself, but also cleaning its metadata—giving clear names and descriptions so that analysts can understand what it all means. With BDD 1.2, data stewards can do exactly that. A deeper Catalog reaches down into attribute names, descriptions, tags, and (new in BDD 1.2) semantic types to making finding relevant data easier than ever. Taken together, these new capabilities allow customers to quickly transform their scattered data assets into an organized, navigable data lake for analysts.

· BDD + Python: Customers no longer have to choose between the ease of use of an integrated visual experience and the freedom of custom code—with BDD 1.2, they can have both. BDD 1.2 offers a new shell mode that allows data scientists to pick up any dataset in the Catalog for custom processing in Python. From here, ingenuity is the limit: data scientists can develop custom transformations, apply advanced algorithms, and perform predictive analysis with the tools they know best. This integration provides the best of both worlds: a shared, analyst-friendly environment for reusable, collaborative data exploration, preparation, and visualization (BDD) plus a dedicated environment for advanced, narrowly tailored processing.

· Streamline + Share. Most big data tools optimize for technical specialists--and they end up freezing out organizations’ most potentially powerful assets: their established teams of business analysts. BDD 1.2 tackles the problem of large-scale collaboration head on, by streamlining the visual data exploration and data transformation experience and by building a notification framework to keep all users on top of what’s going on. Combined with the enhanced data shaping and data curation capabilities, these enhancements open up the data lake to more than just data scientists.

· Speed + Scale. A combination of open-source maturation and Oracle engineering makes 1.2 the fastest, most scalable BDD yet. Users can acquire and transform data significantly faster than in BDD 1.1 and at significantly higher scale and concurrency. BDD 1.2 also includes the ability to store Dgraph indexes in HDFS, eliminating the need for shared NFS and providing a native, high-performance option for in-cluster deployments.

These are just the highlights—check out BDD 1.2 for yourself to see Oracle’s strategic solution for giving organizations a competitive advantage through big data analytics.

Tuesday Mar 22, 2016

May the Source be with You

The data source, that is.

We’ve all been there. Your manager asks you for analysis on how a particular project is progressing, and, being the tech savvy business manager you are, you head over to your IT-managed system that has the data sets neatly modeled for you. Even to the most seasoned professional, this is no small task – a fair amount of querying, filtering, and dragging and dropping is to be expected. But, you’ve done it all before, so you get to it, and in due course, your analysis is done!

Not so fast.

Your manager sends over a couple spreadsheets, and asks you to add it to your analysis. Suddenly, this previously manageable task turns into a “please do this seemingly impossible thing that just popped into my head and get it back to me in the next hour” type of request, and now you are left scrambling on how to get this done.

Luckily, your department often has to mash up data from different sources before. In fact, leveraging multiple data sources is a pretty common occurrence, and your company has invested heavily in making sure accessing data is not an obstacle to performing an analysis. But what about when you have a personal data source you would like to add (or your boss’s ultra-important Excel file)? Tapping into the system to access modeled data is one thing – but mashing it up with your personal files is quite another.

Enter Oracle Data Visualization.

Blending these two types of datasets is key to having a complete view of the business. On a day-to-day basis, we gather information to measure, manage, and respond to what is happening in our business. And the reality is, we need to have access to multiple database sources and the have the ability to blend that data with our personal files. In so many of the projects that we do, we throw data points onto a spreadsheet for one reason or another, and this dataset is no less important than the modeled data we get from our friends in IT. No data should get left behind simply because it’s difficult to use.

So the next time your boss gives you a 1-hour turnaround time for an analysis – remember, don’t sweat the small stuff.

Try a free 30 day trial of Oracle Data Visualization Cloud Service here.

Pati Nishiyama

Monday Mar 07, 2016

Moving Towards Analytic-Driven HR: Does HR Need Better Analytics?

With tight competition to maintain the strongest employee pool, today’s Chief HR Officers and HR leaders are constantly asking themselves strategic questions. Here are a few:

  • Do we have effective employees? How do we analyze our talent: Its ability to produce? Its teamwork capability? Its financial ROI?... • Are we paying employees a competitive wage? What are the risks of losing valued employees to the competition? How are similar companies doing…by industry? …by company size? …within our geography? 
  • Are we hiring the best employees available? Where can the best candidates be found? What is the perception of my company by the candidate pool? What are the employment conditions and perks expected by today’s candidates?  

Undoubtedly, the more relevant information that a CHRO and HR leaders have available, the clearer the picture of the company’s ability to attract, develop, and maintain the best employees becomes. It’s also evident that some of this information isn’t available from the traditional HR data sources.

When comparing the analytics needs of  HR leaders with what’s available to them I find that, for many HR departments: 

  • The information in the data repository is limited and often the data isn’t available from a single system. 
  • Use of analytic and automated reporting is scarce. In fact, many reports are manually compiled in Excel and, often, the information is stale even before it’s produced. 
  •  Reports tend to focus on history and summarize factual events. How is an HR leader to answer the strategic questions, today, with the information that is readily available? 
Clearly, CHROs and HR leaders need to gather information from not only traditional sources, but from a growing variety of non-traditional sources. In a subsequent article, I will expand on the informational demands being placed on CHROs and HR leaders.

Monday Feb 22, 2016

Oracle’s Perspective on the Analytics Market

Rich Clayton | Vice President 
Oracle Business Analytics Product Group

Amidst the constantly evolving analytics market, there has been a fundamental shift from IT leading the charge to pursue analytics initiatives, to one where the business and IT share in this decision. There’s no doubt that analytics has become strategic for most organizations today, and as such has introduced a new wave of both new consumers and new expectations.

The workforce is changing, and with that change brings a new way to work. Gone are the days where training manuals are commonplace in the office – today’s workforce expects to get up and running quickly with an intuitive interface. But it doesn’t end there. While speed and simplicity are key, business leaders still have high expectations around data quality and security. A centralized analytics platform where IT plays a pivotal role is still a fundamental part of any analytics strategy. The combination of both business led and IT led initiatives is the sweet spot for innovation.

We didn’t just come up with this on our own - customers have spoken, and Oracle has listened. Here are some of the top analytic themes our customers want.

A Path to the Cloud. Cloud and analytics are a perfect match. Cloud is about speed and innovation, not just different economics or a pay-as-you-go subscription model. We’re seeing customers go from purchase to production in 17 days – for 700 users. Most importantly, we are seeing cloud change the dynamic between business leaders and IT professionals and how they work. Keep in mind, this doesn’t assume all data lives in the cloud. Our strategy is to provide a best-in-class hybrid approach that extends current infrastructure investments and data security with the speed and innovation of the public cloud.

Data visualization. Business people want to see the signals before their competitors. They want to mashup personal, corporate and big data and quickly understand the value of their data, share their data story with their colleagues, and do it all in a matter of minutes. Data visualization is key to the simplicity of this, and therefore needs to be embedded into every analytic system from mobile to desktop. From department to enterprise. From on-premises to cloud. What’s essential is continuity of experience for fueling the analytical productivity that is expected today.

Data discovery for all. This is no longer a task for data scientists but for anyone trying to experiment and test the value of their data. Discovery is not just a better BI, it’s a whole new method for enriching data, exploring patterns and building perspectives with qualitative and quantitative data.

Mobile. We’ve been using analytics on our devices for some time, but now the market demands a location-aware “assistant.” Our mobile experience not only supports voice recognition and mobile authoring, but also looks at your digital footprint to present insights relevant to your upcoming meetings.

To support these trends Oracle has designed, built and launched several new offerings in the last few years, such as Oracle BI Cloud Service, Oracle Data Visualization Cloud Service, Oracle Big Data Discovery and Oracle BI Mobile. Most importantly, these new business-led solutions are tightly integrated with Oracle’s enterprise analytics platform – Oracle Business Intelligence Suite.

Our strategy ensures consistency of experience and portability of analytic content across business-led and IT-led analytics initiatives, whether cloud or on-premises. And we strongly believe that the combination of these two modes of analytics working together is what will create sustainable competitive advantage for our customers.

Monday Feb 08, 2016

Is Human Resources a Cost Center or a Value-Added Partner?

Call it what you may – “Human Resources”, “People Management”, “Human Capital Management” - is a new name, by itself, sufficient to validate its importance to the organization? Traditional business areas have changed the perception that they are “cost centers” and have elevated themselves to “value-added business partners.” For example, the traditional finance operational functions – keeping the books, ensuring compliance, etc. – are still important for an organization’s survival and must continue. Those in finance roles, however, have also realized that they are in a unique position to advise and to partner with business executives to drive and more closely support the organization’s strategy. To accomplish this, they have had to prove that they are able to look beyond the organization as a series of financial transactions to emphasize how finance and financial decisions support the objectives of the business entity. Surely, traditional human resource roles – hiring, benefits and payroll management - are necessary for an organization to function, but these are operational tasks. What does an HR department need to do to attain an elevated organizational role? In her 2015 article, The New Roles of the Human Resources Professional, Susan M. Heathfield observed, “It's not enough to ask for a seat at the executive table; HR people will have to prove they have the business savvy necessary to sit there.” So how does HR transform itself into a value-added partner of the organization? The answer lies in using robust Talent Analytics to Drive Performance. Analytics-driven HR:

• Delivers more precise visibility to predict and plan organizational performance
• Permits cross-functional insights to better understand, develop and maintain a productive workforce
• Measures and shapes HR’s contribution to the business

The next question, then, is how do HR departments acquire Analytics-driven capabilities? In subsequent posts, I will discuss what it takes to move to an analytic-driven HR organization, how HR can always be one step ahead, and how to build and maintain top talent.


We're taking the pulse of the Business Intelligence and Analytics market based on insights and experiences with colleagues, customers,and partners.


« July 2016