I'd like to thank all the organizers, host and presenter for their time and effort yesterday. I am excited to see a growing tech meetup community which I hope to participate fully in. I enjoyed meeting people, and Peter's presentation.
I'd like to provide feedback, which I hope is constructive for future events.
I thought the session was slightly misleading in the advertising. I had expected to hear about real experience of big data analysis using spark and specifically on machine learning algorithms - perhaps touching on topics like infrastructure tuning, runtime performance on various workloads, partitioning strategies or which primitives of spark are suited to which statistical analyses etc. I think the presentation was more geared at watching someone go through some "hello world" level examples and without real experience of "big" data - a 3 node temporary cluster built from desktops is rather early in a big data infrastructure adoption path.
It is totally necessary to have more introductory level meetups but it wasn't quite what I had expected. Perhaps being a little clearer on the content would help in expectations? - "Running your first Spark job" would have perhaps been a more suitable title, which might have attracted less people but might have resulted in some good interactions - it could even have been a hands on tutorial.
I have a couple of minor suggestions for organizers of future events:
-
I think it would be good to stick to schedule. The first session was 1.5hrs, when I read somewhere the target was to be 45min, break, 45min and then discussion. I'm afraid I had to leave at the break so I don't know what the second session was like but I would have enjoyed being around for the interaction and discussion part.
-
I found it difficult to hear at the back and would recommend organizers to consider that in future.
I'd like to reiterate my thanks to all involved and to Peter for taking the time and effort to present. I hope this all comes across in the manner intended - constructive feedback and not just criticism. Thanks Peter and good luck with your research, and thanks Vladimir for driving all this.
[I'll try and find out if we can provide a venue for future around the Zoologisk Museum. If anyone were interested I'd be happy to talk our lessons learnt over the past 6 years on on adopting and building applications on Hadoop (HBase, Oozie, Hive, SOLR cloud, Hue and considerations of Spark, Storm and Impala). We're involved in a open data network for biodiversity information exchange and run real time indexing and mapping of biodiversity data http://www.gbif.org/occurrence].
Please do not comment on this gist - comments are to be on http://www.meetup.com/Big-Data-Denmark/events/225580411/