Southern Data Science Conference

Atlanta, Georgia – April 12, 2019

The Southern Data Science Conference was a super informative conference with a focus on state of the art data science coupled with opportunities to network and to use data science for good.  The conference follows a unique pattern — all presentations are done in a single large room with presenters quickly changing off.  This method avoids a lengthy time for participants to go from room to room.  I was pleasantly surprised by how well the approach worked.

The conference draws on speakers from across the country – not just the Atlanta area.  Organizations such as:  Google, Amazon, Lucidworks, Netflix, Microsoft, SalesForce, Hitachi, Target and Vertica provided speakers.

TensorFlow was discussed in multiple sessions and I became excited the TensorFlow 2.0 is available.  This is a conference where actual code is presented – in that case Python rules as the universal language for showing data science examples.

Presentations by Netflix and Pandora gave an inside view on the creation of recommendation engines.  Both of these firms have different recommendation engines for different parts of their websites.  Each recommendation engine then calls upon dozens of recommendation algorithms to generation recommendation and then a ranking function selects the best recommendation.  Pandora also employs a massive music knowledgebase.  HomeDepot, well represented at the conference, indicated that it also employs numerous recommendations engines based on product categories and other functions.  These concepts are helping in my efforts to improve the recommendation engines that I have developed.

The primary background of data scientists is math and statistics.  Earlier I thought that DBAs, database developers, BI analysts would be the source of data scientists.  Instead, I see that this group is now the source of “Data Engineers” who build data pipelines to satisfy the needs of the Data Scientists.

In conclusion, this is a worth while conference for those who want to improve data science acumen.  I learned a lot, took over 40 pages of notes and made many valuable contacts.