Looking Back At Oracle OpenWorld
- 16th October 2017
- Oracle BI - General
- Tony Heljula
It was another fun week at OpenWorld, with San Francisco as beautifully-sunny as always.
From talking to people prior to the conference, you got the feeling that people were a bit concerned that this year would be all about cloud and nothing else! Apart from the odd exception, I don’t think that was the case at all.
Even when a presentation had a cloud focus, it seemed at least that we were now able to focus on innovations in the cloud (such as the new Oracle Autonomous Data Warehouse) rather than what products were in the cloud.
Moving Straight Onto Analytics
I thought that this year’s analytics roadmap presentation was excellent - one of the best ever in fact.
Last year, the analytics roadmap sessions seemed to spend all their time trying to explain the differences between BICS and OBIEE. This year thankfully, it was all about new features. One could argue that there were actually too many new features to take in.
The focus was naturally almost entirely on Oracle Data Visualization (Data Viz). In some ways, this was a bit unfortunate as we don’t want Oracle to forget that 99.9% of its analytics’ users access content via Dashboards. We do still need some innovation with OBIEE, if nothing else, to integrate Data Viz into the Dashboards (much like the way BI Publisher was integrated right from the beginning).
That aside, the innovation and rate of change within Data Viz is pretty huge. It is clear that Oracle are not trying to man-mark every Tableau feature (Gartner once suggested that Data Viz was simply playing catch-up). Instead, Oracle are actually giving Data Viz its own unique personality - especially in the areas of machine learning, data science and predictive analytics.
There will be a new ‘machine learning’ component of Data Viz - to deliver the type of capabilities you currently can only really get with Oracle Data-Mining - but the features go beyond that. Whilst I am sure ‘version 1’ will have some limitations, this is the just the type of capability that we have been needing for a long time. It makes open-source, predictive analytics much more accessible to the masses (you don’t have to learn how to write R scripts etc.).
Highlights Surrounding What’s To Come
Here are some of the highlights of what is potentially coming over the next 3-12 months:
- Machine Learning: Build and test predictive models within Data Viz (based on Apache Spark algorithms)
- Natural Language Generation: Providing automated narratives / commentary to go alongside your visualisations to explain the data content
- RTD Features: Model management, online scoring, A/B testing
- Data Flows: ability to build, test, training predictive models. Streaming data flows. Data flow job monitoring
- Explain: One-click predictive/prescriptive analytics. For example: what contributes to my large employee churn?
- Bare Metal Cloud: Options for hosting in the BMC
- BI Publisher: BIP ‘pixel-perfect’ reporting will be available in the Oracle Analytics Cloud. BIP reports will be able to use Data Viz data sources for its reports
- Essbase Cube Creation: Build Essbase cubes for fast/interactive visualisations
- Data Preparation: Intelligent recommendations for data blending
- Packaged Applications: New class of DV applications for CX data sources
If you have previously produced any documentation or presentations on Data Viz, then you’ll no doubt be pleased to hear that the UI is changing altogether in the next release – so you’ll have to update all those screenshots once again (haha J!)
Other features and components that got good mentions were Oracle Analytics Cloud (OAC) and the new mobile apps (called ‘Day-by-Day’ and ‘Synopsis’).
For the second year running, it didn’t surprise me that there was hardly a mention of Oracle BI Applications throughout the whole conference. Whilst nobody has directly said it, I get the feeling that BI Apps is a very big ship that cannot cope with the rapid rate of change that the cloud brings - and the variety of different cloud sources that are now available.
BI Apps has also historically relied on having direct access to source databases, which no longer happens with cloud hosted apps (where data is typically extracted via web services).
So whilst BI Apps has been certified on the Oracle Analytics Cloud (OAC), I wonder if we should no longer expect any BI Apps releases in future?
Smaller, More Agile Applications
For packaged applications, it seems the focus from now on will be on producing smaller and more agile applications in Data Viz that can connect direct to source cloud applications (e.g. CX Cloud), using OTBI as its interface.
OTBI is essentially an instance of Oracle BI hosted within the cloud application, for operational reporting purposes. Data Viz can extract data from OTBI either through issuing Logical SQL against the OTBI subject areas, or by sourcing data from its pre-built analyses.
In summary, within the Oracle analytics’ space, the future is definitely with Oracle Data Viz. It is rapidly turning into an incredible product with lots of wow-factor. But my main hope is that Oracle don’t forget about the importance of dashboards and dashboard prompts to the 99.9% of end users who use them daily.
Data Viz, in its current form, does not replace the need for OBIEE dashboards. So dashboards cannot become legacy, even if they aren’t as exciting as machine learning.
A great roadmap though - that’s what we were there for.