I recently attended the ODSC EU 2017 a two days conference held in the amazing city of London. I was excited to have the opportunity to join this event, with more than 1500 attends, 75 speakers and 28 workshops. The conference allowed attendees to connect with some of the most innovative people and ideas in the world of data science, while learning from core practitioners and contributors.

As data science touches nearly every area of our lives, the event covered many aspects like Open Data Science, Machine Learning, Artificial Intelligence, Quant Finance and Research. Despite the wide range of topics offered, ODSC 2017 delivered interesting and novel contents in an enjoyable atmosphere.

Open Data Conference

 

In the last few years, some fields have started to gain more and more attention within the community and industry of Data Science.
This year one of the hot topic was Quant Finance and it was one of the most discussed subject: with its own workshops and lot of presentations, it was a topic that has been taken up and emphasized almost by every speaker.

Word Cloud Open Data Conference

Second to Quant Finance, was Machine Learning Automation Tools. In business environment, the need to have some magical tools acting like a black box that extracts automatically useful information from your data is becoming stronger and stronger.
Of course these services don’t want to replace the Data Scientist figure, rather they are trying to support the work of this role providing some shortcuts to save up a lot of time when setting up different environments, algorithms, processing data, etc..

 

Flow

Here a list of the 3 most interesting talks which was well worth attending:

 

  • Automating Drug Target Discovery with Machine Learning (Enrico Ferrero)

In my opinion, it was the most interesting talk among ODSC EU 2017. Enrico Ferrero, computational biologist at GSK, brought out how drug target identification is a hot challenge in pharmaceutical industry with a high failure rate. Together with his team he applied Machine Learning Algorithm to classify a successful target drug, with an accuracy of 71%. This may significantly reduce the initial search space avoiding huge failure and reducing the cost and developing time of new drugs.

  • Astronomical Trading (Tanya Sandoval)

It was a nice and fun talk held by Tanya Sandoval, Data Scientist in a trading firm. She made a nice introduction to quantitative finance highlighting both its limitations and strength through practical examples such deploying a trading strategy according to the sun and moon astronomical positions.

  • The Magic of Dimensionality Reduction (Alex Peattie)

CTO at Peg.Co Alex Peattie held a beautiful speech about one of the most crucial tools of any Data Scientist, Dimensionality Reductions. It was a complete walkthrough over 4 popular different techniques (Random Projection, PCA, IsoMap, t-SNE). Explaining this powerful methods using funny example (such as Tinder Automated Swipe) he showed the mathematical assumptions behind this method as easy and unnecessary 🙂
For Roialty, being an innovative startup is a key strategy to engage with the tech community, share experiences, get feedback and stay up to date with the last researches.

With its access to a huge amount of social data, Roialty integrates and deploys Machine Learning pipelines to model user behavior, to automatically cluster your audience and provide useful real-time insights.

The following two tabs change content below.

Marco Edemanti

Graduated from Polytechnic of Milano in Computer Science with a thesis on hybrid recommender system. His interests are mainly focused on Machine Learning and Artificial Intelligence fields. He tries to undercover the hidden pattern in the social media and web data. In the free time he loves running and skiing. He joined Roialty in the early 2017 as a Data Scientist.