The combination of short keynotes and panel discussions meant we got to hear lots of perspectives about a lot of interesting projects and programs; and speakers really had to focus to get their points across against the clock.

Computer Business Review wrote it up :

Christian Prokopp, a Principal Consultant here at Big Data Partnership, was in Bangladesh on holiday, and he took the opportunity to hook up with some local Big Data activists.  He quickly found a host of local enthusiasts: academics, entrepreneurs, data scientists and just plain interested tech-folk, and he was immediately invited to a series of well-attended events. The write-ups of Christian’s talks are .

I guess it just goes to show that data knows no boundaries.

But all in a good cause; all the participants who insisted on sharing their taste in seasonal knit-ware also had so share a little of their monetary good fortune with Save the Children.

Thanks everyone. Merry Christmas to all friends of Big Data Partnership. Here’s to a successful 2015!

We’ll be starting the project in 2015, but we announced it this week, and i think our interest is matched, if the press reaction is anything to go by.  had the story first, but others picked it up. Let’s hope the data really does answer this question. But that’s the beauty of machine learning, if there are patterns in the data it will find them, even though they could be patterns far too complex and infrequent for any clinician, even the most skilled and experienced, to recognise unaided.

Triggar uses Cassandra, Spark, Elasticsearch and some very effective machine learning so eCommerce site operators can dynamically optimise site behaviour depending on the profiles of visitors, and how they are progressing through the site. Triggar was initially developed for Postcode Anywhere’s own site, but they  realised the value it could deliver for any suitable site.

That’s when Postcode Anywhere spoke to Big Data Partnership: because Postcode knew they didn’t have the Big Data skills necessary to scale Triggar up.

We delivered those skills. First we conducted what we call an Architecture Discovery workshop to nail the scaling problem. Then we put into place a Development Support engagement, so we could we supply the right skills at the right time to help Postcode Anywhere’s team to crack on with development. It was as part of this engagement that we introduced the data science skills that made a key breakthrough: Triggar would have only understood pre-programmed static routes through the site (what Triggar calls ‘Games’). But now Triggar discovers Games dynamically by using machine learning to recognise patterns in user behaviours and then goes looking for those patterns as new and returning visitors hit the site.

With this evolving library of Games, Triggar can evaluate the probability that any visitor will reach conversion; Triggar calls the visitor a Player (of a Game) and the probability is their Mood. Triggar can then cause dynamic changes in the site to attempt to improve Players’ Moods. It’s really very cool stuff.

Jamie Turner, Postcode Anywhere CTO spoke about it at the on 4th December in London.

There’s on the Postcode site about the app, so i won’t repeat that here. But if you are into Spark Streaming and Cassandra, then do come along.

Slide deck from the talk will be linked in a comment below here, i hope, when it’s available, for those who can’t make it.

It makes , and focuses on the business and not just the technology – which is merely the means to delivering value to the business.

You can read what Christian shared about Hadoop, Big Data adoption challenges, alternative and future trends here:

But one critical aspect of providing support for scale-out technologies like Apache Cassandra is that it’s pretty hard to knock a cluster over. With good configuration and sound management policies there’s no single point of failure. That doesn’t mean there are no bugs in the software; it just means there’s less chance of a ‘P1’ Ticket: production system unavailable.

That said, we did get a P1 on a client’s Apache Cassandra cluster the other week.

And what happened next was pretty interesting. I don’t generally go in for sporting analogies, but this was definitely a full-court press. The guy who picked up the ticket was about to finish for the day, but he couldn’t go until the situation was stabilised, and in between getting a remote debug connection, downloading source code, searching bug databases, trawling forum posts, he had to brief management. Pretty soon we were talking at all levels with the client and our guy on rota had an ah-hoc team of colleagues helping out.

This kicked off about 5 pm on a Friday. By 10 pm the team, including CEO – who had been on the phone and IM with the client CIO all evening, was in the pub. The situation was recovered, the client cluster up and running with a workaround and the necessary patch identified.

But there were still several phones on the table, all targets for the ticket alert – just in case another P1 arrived. It didn’t.

After this was kayaking in the sea, which was fun for everyone who didn’t end up underwater. That evening after dinner we completed the unconventional triathlon by walking into the town centre and drinking beer and Stimulants (vodka, Kahlua, Tia Maria, sugar syrup and a double espresso). Surprisingly, everyone was conscious on the Sunday morning for archery and golf, although the number of un-popped balloons left on the archery targets suggested otherwise. The afternoon was left for relaxing in the hotel’s spa, sitting outside and reading, or alternatively completing a centurion (not a scheduled or obligatory activity). Following an evening meal in the hotel and a few more drinks, we went to bed. The next morning we had an early flight back to Gatwick and returned to work refreshed.

Having joined Big Data Partnership the same week as the trip, I was eager to get to know everyone better. Everyone had already made me feel very welcome, so I wasn’t apprehensive about the trip, and knew I’d have fun. However I was surprised at just how friendly everyone is. I knew almost immediately that I’d fit in well at Big Data Partnership, and felt that the weekend was a great accelerator for becoming part of the team.