I’ve attended Oracle Big Data and Analytics Summit 2016, Istanbul last week. I would like to share my impressions and the key points of the event.
The event is located at Esma Sultan as usual. The building is situated in Ortakoy and it has a brilliant bosphorus panaroma. I am familiar to the place from the Turkish journalist Guneri Civaoglu‘s weekly tv programme Transparent Room.
The panel has started with the talk of Paul Sonderegger. He pointed out the tendency of data reproducement. In other words, data produce more data. Running different machine learning algorithms on your big data produces data again. Moreover, falling the hardware prices deploys Internet of Things in the center of the life. Sensors of IoT devices record daily data recurrently. He also emphasized the data equality concept. The data could be stored at relational databases, data reservoirs (nosql databases) or data dumps (even log traces). They are all valuable, and you should be able to process all kind of data.
Then, new marketing strategies are discussed. Formerly, companies have products and they look for customers. Today, corporations have customers, and they look for the right product for each customer. For example, you plan to buy motorcycle. If this potentiality could be predicted, then you should be faced with the vecihle loan advertisements. This concept is called as Next Best Action. I strongly recommend you to watch Rob Walker‘s keynote if you interested in this concept.
After then, Pelin Ozbozkurt talked about real world examples for rare events and customer segmentation applications. The success of the customer segmentation cannot be measured like supervised learning applications.
Thereafter, she mentinod a funny example. Clustering two males who are born in 1948, grew up in UK, married twice, have two children, successful, wealty, celebrity and like dogs and Alps, in same segment could be surprising. They are Prince Charles and Rocker Ozzy Ozbourne. You cannot expect them to have same purchase behaviour. Herein, big data provide to group and cluster the same rare events.
Finally, she quoted the statement by Peter Norving (Director of Research, Google) that indicates “We don’t have better algorithms, we just have more data”. Herein, I would like to quote a similar statement by Andrew Ng (Assoc. Prof., Stanford) that indicating “It’s not who has the best algorithm that wins. It’s who has the most data”.
These are the notes I’ve taken in this year’s Big Data and Analytics Event. I’d look forward to upcoming activities of Oracle Big Data group.