I attended the Oracle User Group Ireland this year on a beautiful sunny Dublin morning. I really love to get away from my desk and see what is going on in Ireland’s Oracle user community.
I was able to catch up with a good few people I know, and bumped into to some unexpectedly, which was a nice surprise. The conference is held at the convention centre, centrally located and a purpose-built conference venue.
The conference was very well organized with a good mix of speakers, sponsors including big names. I found it hard to choose sometimes what session to attend, as there were lots of interesting stuff going on.
Thanks to OUG for organizing it!
I was running a bit late so got the benefit of a very short registration queue, and just made it in to hear the keynote of Jon Paul the Oracle regional director. The interesting takeaway for me was that Oracle are selling direct to Chief Marketing Officers not, rather than CIO’s. Seems like the business are spending directly without involving IT.
Big Data Analytics with Oracle – Uli Bethke (http://sonra.io/)
So my first session was Big Data Session with Uli Bethke of Sonra (Sonra is the Irish for data). He explained Big Data in layman terms and gave some really interesting use cases stretching from tank production in WWII to self-driving cars in Silicon Valley. Uli also talked about the Amazon Anticipatory Shipping program as an example.
Uli talked about how to integrate Big Data Analytics into the enterprise data architecture, and there was some good questions and answers on how exactly this might be done in various real-life situations.
In terms of getting started he recommends to Start Small; Define Use Case; Use Market cloud based services; use experienced Data Scientists and beware Taleb’s Black swans; that is to understand the difference between signal and noise.
Oracle Fusion Financials – The Next Step in Finance Evolution – Tracy Bishop Oracle
Next on my list was the Fusion Financials. This is an area I’m keeping a close eye on as an eventual replacement for Oracle EBS. A friend of mine has already completed the Fusion implementers bootcamp so I wanted to find out a bit more on Fusion and if I should be getting trained up now; more on that later.
Tracey Bishop explained three headline reasons for Fusion; Cloud, Integrated Intelligence and Collaborative Tools to satisfy three design goals of Fusion; what do I need to know, what do I need to do, and who can help me.
It has to be said the product looks great based on the demo. It seems like Oracle tried to incorporate business analysis and presentation tasks currently performed outside ERP into the feature set. In addition Instant Messaging is also part of Fusion so the history of previous issues and resolution is stored in the ERP, not in mails, external files, IM etc.
I asked what the customer take-up is of Fusion Financials is in Europe and it seems there is still only one implementation in northern Europe. Cloud-based CRM and HCM are far ahead of Fusion financials for the moment in terms of customer base. However Tracy did say there are clients evaluating the product for future implementation. All in all I wont be booking my implementers course yet, but I will be keeping a close eye on the communications on Fusion.
Sentiment Analysis using Oracle Analytics – Brendan Tierney http://www.oralytics.com/
I am very interested in Sentiment Analysis and always try to follow what Brendan is doing in this area. This also relates to Oracle’s statement that they are now talking to Marketing departments as marketers are definitely very interested in this type of data.
According to Brendan Sentiment Analysis using Data Mining is about explaining the past to predict the future, it can be used to evaluate the sentiment of a document; blog post, forum, twitter feed etc and decide if the sentiment is positive or negative.
Each sentence in the document is broken down into component parts using a technique called tokenisation, and the ones that do not indicate opinion are removed. The parts that do indicate positivity or negativity are compared to a previously labelled list, which requires human effort to compile.
Brendan based his presentation on a live demo using Oracle Data Miner and Oracle Advanced Analytics. In 30 minutes he created the framework to examine sentiment based on a feed of movie reviews.it was very smooth progress through the demo using Data Miner, and he promises an OTN blog post with the detailed steps.
The practical applications mentioned by Brendan were a telecoms company that uses sentiment analysis to predict churn rates and timing. By analyzing a customers interactions with the company, and their posts on public sites like Twitter for instance, they can predict if and when they might leave.
Some things to watch out for are include labelling of positive and negative compare statements, and for instance that this labelling will not understand sarcasm. So a statement like “The movie was great – not” would be interpreted as positive.
I also talked to Brendan about the publication of his upcoming Oracle Press book ” Predictive Analytics Using Oracle Data Miner: Develop and use Data Mining Models in ODM, SQL and PL/SQL”. It contains 20 chapters of information to build next-generation predictive analytics applications and should be the reference book for this area.
So overall I had a great day, learned some really useful stuff, and have some food for thought in terms of my future career plans. Looking forward to next opportunity for meeting up, I think there is a Special Interest Group in autumn so I will try to attend that.