Business data paradise - we visited Strata Data conference in London

Deep learning, AI, Big Data, Cloud, Business Data Analysis and Data Visualisation - these were the most important buzzwords of Strata Data conference, which was organised in London from 22 to 25 May. Here’s the most important highlights of the conference from our point of view.

The theme of data-driven technology drew a huge amount of fans to the event. It is almost obligatory for data engineers and data scientists to attend one of the strata Data events which is now held in a number of cities worldwide. The speeches and workshops of the conference, which was organised by O’Reilly Media and Cloudera, were about Big Data, Cloud services and Artificial Intelligence, which are all closely linked together and make up new and complex service packages today.

Because what does a Data Analysis strategy look like in practice nowadays? The data which comes from more sources, is totally unstructured and has more than one different types (Big Data) must be stored on a platform (Cloud) which is able to handle the amount of data which grows constantly. The Cloud also needs to be able to deal with the algorithmic processes connected to the data (Artificial Intelligence). The three elements are connected together and make up a unique system which can also be complemented by other plugins and modules.

At the conference, it was Artifical Intelligence themed speeches (complemented with Machine Learning and Deep Learning) that drew the most attention and interest. Many big companies presented their developments, AI devices and modules. The most important expectation of such a solution is that it should fit into the system as a trustworthy, flexible and customisable block - and with the help of algorithms, it should also process the huge amount of data in the shortest time possible. In the competition of such solutions, the most popular framework was the Google TensorFlow - but there were also other big names in the game, such as Microsoft, Intel, Dell and Nvidia. At the conference, we had the opportunity to try TensorFlow ourselves, too - we are planning a blog post about our personal and professional experience in connection with it.

As the biggest Cloud service providers, Google, Amazon, Microsoft and Cloudera presented their newest solutions, which are able to handle big amounts of data quickly. To show us the size of their databases, they told us that smart cars developed by Uber collect 1 TB of data each day. This also means that with the increase of the amount of cars themselves, they will generate an unbelievable amount of data in the future. And these databases are usually unstructured and not suitable for automatic processing. That’s why Business Data Analysis strategies must be completed with data wrangling. The aim of this activity is to turn unstructured Big Data into information which can be interpreted and automatically processed and which is also free from errors and duplicated data.

And what are the business possibilities in the Big Data/Cloud/AI trio? Uber, for example, tries to take advantage of data analysis on all fields possible: with the help of its algorithms, a car can drive to a passenger before they even ordered a ride - based on prediction. And then, when the car had arrived, the passenger can make sure they have the right car with a face recognition module. This way, passenger safety greatly improves as well.

All in all: we have the building blocks. Now it’s up to the builder how they use them: in what order, what relation and for what purposes. The importance of data is obvious and cannot be ignored, so we’re looking forward to apply the technologies we learned at Strata Data in our own development works, too.