Last week was the fifth Berlin Buzzwords conference, held in the lively Prenzlauer Berg area of the city, and fellow liiper Lukas Kahwe Smith and I were there. Berlin Buzzwords has three tracks on data analytics and processing at scale: search, store and scale. As well as finding out about the big data topics of the day, I went there hoping to find out some tips and tricks for Solr and Neo4j, which I've been interested in since Philipp Küng recently gave a techtalk on it at Liip.

Sunday: Barcamp

I travelled to Berlin during the day on Sunday, to be there on time for the Barcamp at the start of Buzzwords. It was a good opportunity to get my wristband and name badge before the crowds appeared on Monday! My favourite talks were Anders Hoff‘s presentation of his generative algorithms in Python (his code is here) and a discussion on how to make Wikipedia's search more useful for power users.

Monday

The catering throughout Buzzwords was impressive, but it was when I got there for breakfast and saw the never-ending Club Mate bottles on the bar that I really knew I was in Berlin.

Felipe Hoffa and Ewa Gasperowicz's talk on Exploring the Notability Gender Gap was maybe my favourite of the whole conference. They combined Google's BigQuery and Maps to look into the question of gender representation in Freebase, and I left excited to try out BigQuery myself.

After lunch I enjoyed Michael Hunger's workshop on Neo4j GraphGists and Michael Kaisser's talk on Geospatial Analysis of Social Media Posts with Elasticsearch.

Tuesday

Tuesday had two very useful presentations for my quest of learning about Solr: Side by side with Elasticsearch and Solr by Rafał Kuć and Radu Gheorghe, and Chris Hofstetter's Hidden Gems: Getting More Out Of Apache Solr ( slides here). Otherwise, the most interesting talk I went to was Breno Faria's on content tagging at ZEIT Online, where tags needed to be relevant for current material and as full as possible for the archives, and neither robots nor humans alone could do a good job. There's now an API for the tagged content, too!

Wednesday: Neo4j meetup

The Berlin Neo4j group had organised a Learn Hack-Day for the day after Buzzwords, with about 50 people attending. I think everyone was happy to learn that Neo4j 2.1 now allows bulk import of data from CSVs, and, fuelled by pizza and more mate soda, we worked on projects ranging from cocktail recipes to political donations (see #graphhackday for examples).

Finally

I had a great time at Berlin Buzzwords: discovering cool new tools, talking about big data, and exploring the city too. I'm looking forward to July, when Liip will return there for OKFest!

Some additional notes from Lukas

I have attended several previous Berlin Buzzwords editions. This time I especially focused on learning more on Cassandra. The fact that they now seem to prefer an SQL-ish interface makes it much more approachable and is just another pointer at how ridiculous the NoSQL acronym is. While PHP seems to be a tabu topic at the conference, I learned via twitter that there is infact even a PDO driver for Cassandra. Other than that I tried to hear about as many different technologies as possible, be it YARN, Redis, ElasticSearch, Solr etc. In general however I felt this edition was a bit too focused on ElasticSearch, Hadoop and Cassandra. That being said, my favorite talk was Britta's talk “Scoring for Human Beings”. Just liked her style and the fact that she both provided some theoretical background combined with useful information that one can quickly apply in the real world.