TDWI Conference 2017

Earlier this month, Richard Roose, one of our senior Sydney consultants, attended the TDWI conference in Anaheim. He gives us a recap of the main topics below.

As you might expect, a broad spectrum of topics were covered, with some of the hotter ones being around:

  • Self-service BI & Analytics
  • Data Science & Advanced Analytics
  • Data Visualisation & Storytelling
  • Modern Data Architectures

Disney was the location and Disney was also the opening keynote with some interesting observations. For them, it’s all about digital engagement and customer experience. 80% data quality is good enough. Their SDLC is a hybrid approach, 80% agile, 20% waterfall. Their predictive analytics is getting email open rates of 75%. They allow customers to self select into marketing channels and they had strong advice to keep analytics in-house for competitive advantage.

When it comes to Self-service BI & Analytics, this has evolved from the “nirvana” originally promised by vendors to something a little more pragmatic. It is now often characterised by personas like “Data Citizen vs. Data Steward vs. Data Scientist” with each having unique requirements for accessing data and in the case of data scientists, data in any form. Data as a Service is fundamental to serving these varying needs. This in turn is seeing a resurgence around data governance, not in the traditional IT sense but as a more business outcome driven activity. The other important note is that data governance is more directly related to Data & Analytics activities as opposed to source system activities (i.e. it’s more about producing useable data assets and less about full data lineage).

Data Science & Advanced Analytics continues to advance across a number of spectrums with a particular focus on Natural Language Processing and bringing semantic meaning to data via machine learning cataloguing. There is a lot of work on image and geo-spatial processing as well. Predictive modelling hasn’t gone away either, it’s more a case of applying “additional layers” of quality data to fine tune the models and make them more accurate. Whilst a hot topic with media and vendors, AI didn’t really get more than a mention as it’s still relatively early in the hype cycle.

Data Visualisation & Storytelling is now really about finding the right people with the skills to tell a compelling (unbiased of course) story. The toolsets are no longer the impediment and many of the techniques (eg. Stephen Few) are becoming quite well known.

Modern Data Architectures, whilst there are varying opinions, the Data Lake saw plenty of attention and this is really being driven by the needs of the likes of the Data Scientist and/or Marketing to access any data in any format at any time. This need is also driving a break from traditional “one size fits all” data modelling (eg. Kimball) to a more hybrid “data cake” approach where layers of data in varying formats/model types and varying degrees of data quality can be joined/sliced to provide insights and correlations. This in turn, is driving the need for business-driven data governance that makes these data assets searchable/accessible. The diagram below highlights some of the concepts:

 

In summary, as always the only constant is change but it looks to be change for the better and in particular, better decision making, which at the end of the day is what Data & Analytics is all about.

Join the conversation

Your email address will not be published. Required fields are marked *

Comments

Post has no comments.