Data Science Economy: Data Driven
2 keynotes / 16 talks / 36+ speakers
Conference Schedule!

17. – 18.05. 2018, @HUB385, Zagreb, Croatia

RLS: Real Life Use Cases and Stories: Lectures describing details about business and technical perspective of concrete real-life use cases based on relevant subject areas (Data Ingestion, Real Time Processing, Machine Learning, Predictive Analytics, Natural Language Processing,  Data Visualization, Artificial Intelligence etc…

CAL: Concepts and Learnings: Lectures related to general advices, experiences and recommendations about logical architecture, technologies, trends, tools, concepts, project organisation, data governance, data protection and beyond.

WNX: What is Next, Trends and Future Development: anything emerging, looking to be disruptive in the field of data science and applications of data science. Looking at the future, showcasing the things that will keep us awake in the months to come.

Day 1, 17.05. Thursday

Conference Day

Here, you attend the predefined sessions and learn from case studies, presented by great speakers.

Conference Registration

08:30 – 09:00 | HUB385
Secret Note: for all early risers, we will have a free coffee (limited quantity, come early)

Welcome to the Conference!

09:00 – 09:10 | HUB385
Welcome to the Data Science Economy Conference!
Donald Brancato, Microfocus

Keynote: Data and Value in a Time of Uncertain Value

09:10 – 09:50 | HUB385 Academy
The news of the past year regards Cambridge Analytica, Facebook, and even GDPR, has mounted a new campaign of interest on just what is ‘data science.’ And for that matter, what is ‘data engineering’? Compound the ‘what is this all about’ with more news on agility, standards, IoT, smart ‘almost anything,’ and the world has heard a lot! Have you heard about ‘dataOps’?
We’ll examine data science, a definition, and perhaps its history and current traditions, patterns. Who the practitioners are and who uses the data, information, or graphic that the data story tells. How ethics and value are, and will continue to be, a topic of debate, rationalization until common contextual patterns evolve.
Vuk Vuković, ORACLUM

Keynote: The polls were wrong! Our predictions were wrong! How can we fix that?

09:50 – 10:30 | HUB385 Academy
Predictions are hard, especially about the future“, so said the famous NY Yankees player Yogi Berra. Why are we so bad at making predictions? We rely on, make, and listen to forecasts every day. Making predictions is a natural response in overcoming the knowledge deficit in a world filled with uncertainty. So if we know that people are prone to biases and overconfidence, can we still trust them to make good predictions? We at Oraclum say – yes we can! In fact we can use the people’s own groupthink biases to improve their predictions. How? We will show how during this keynote!

Coffee Break

10:30 – 10:50 | HUB385
Leonardo deMarchi, Badoo

Keynote: Integrating Machine Learning into your Business

10:50 – 11:30 | HUB385 Academy
Presentation will showcase some of the best practices to integrate data science into businesses. We will cover all the main steps to achieve the above, from creating a team to deliver business value and become a data driven organisation.
Chris Roche

Keynote: Delivering a collaborative Data Science platform for Clinical Research

11:30 – 12:10 | HUB385 Academy
Execution is a powerful thing. Generating great ideas is easy but a great idea without execution results in zero revenue for your business and in the clinical setting zero patient impact. Collaboration and data are acknowledged as disruptive forces that can transform clinical research yet the current fragmented system acts as an anchor to the past challenging any attempts to change. Having built and delivered a collaborative data science platform for clinical and biomedical research, in this keynote, Chris will share his learnings on how to accelerate a data science-based platform as a service business and discuss some of the research use cases that took advantage of the platform
DSE 2018

Panel Theme: How to become Data Scientist in Croatia?

12:10 – 13:00 | HUB385 Academy

Panel Participants:
Vuk Vuković, ORACLUM
Leo Mršić, Algebra Lab
Stjepan Pavlek, KOIOS
Bojana Dalbelo Bašić, FER

We will hear from the first hand what some major stakeholders think about data science education and what we can expect in the future of our efforts to educate new generation of data scientists.


13:00 – 13:50 | HUB385 Aleksander Pivk, SAS

RLS1 Paradise Found: Discover Your Best City to Live using Machine Learning

13:50 – 14:20 | HUB385 Academy Room A
Have you already found your personal paradise? What country in the world is the best place for living and working? SAS platform knows the answer.
Stjepan Pavlek, KOIOS

CAL1 Data Scientist To Be

13:50 – 14:20 | HUB385 Academy Room B
Is Data Science new step in BI evolution? Who is Data Scientist today, what challenges does (s)he has to overcome and what are must-have skills for that? What is the future of today’s Data Scientist and what is Data Scientist of the future?
Ratko Mutavdžić, Microsoft

WNX1 Quantum Computing and future of Data Science

13:50 – 14:20 | HUB385 Academy Room C
Quantum Computing will be a massive breakthrough for data science and science in general. What will change? What problems we were not able to solve so far, and Quantum will give us the opportunity to do that? Learn about current efforts in Quantum and what will happen in near future.
Violeta Misheva, ABN AMRO

RLS2 Banks are ready for Big Data Revolution

14:20 – 14:50 | HUB385 Academy Room A
Banks are more and more vigorously embracing the Big Data Revolution. This talk highlights some use cases developed within ABN AMRO bank in the Netherlands. It deep dives into a proof of concept developed for the domain of mortgages and determining eligibility. Machine learning proves valuable even when regulation does not permit automation.
Ognjen Zelenbabić, Content Insight

CAL2 Data Science in Media Industry

14:20 – 14:50 | HUB385 Academy Room B
In the recent years we have been witnessing decrease of media quality all around the globe. Fake news and clickbait headlines are deterring readers and lowering their interest in the articles on the web. This in turn is costing media industry a lot of money and pushes media companies to search for new business models while the root of the problem is untouched.
I will show you how we can tackle with this problem and thus restore faith in the media bringing quality content back to the readers. We will also touch the subject of communicating complex results in a simplified and meaningful way so that everyone can understand “science“ firsthand.
Marijan Bracic, Poslovna inteligencija

WNX2 Data Privacy and Monetization in the age of GDPR

14:20 – 14:50 | HUB385 Academy Room C
European union has enforced GDPR, a regulation that is standardizing data privacy principles for all organizations processing EU citizens personal data. The idea is to have the same rules when it comes to protecting EU citizens rights and to stop the uncontrolled exploitation of personal data.

RLS3 Forecasting consumption using predictive modelling

14:50 – 15:20 | HUB385 Academy Room A
Showcase of the real-life scenario where we implemented predictive model for gas consumption forecasting based on internal and external datasets. Primary goal was to provide accurate predictions on hourly level enabling client to optimize ordering and distribution process. (Case Study: Gradska plinara Zagreb)
Gorjan Agačević, AMODO

CAL3 Why would anyone share anything personal with you?

14:50 – 15:20 | HUB385 Academy Room B
We talk a lot about sorting, analysing and generally making sense of data. Many enterprises have big plans about what they will do with all the new data they plan to obtain from their customers and then get unpleasantly surprised when their customers say: “No, you can’t have this.“ With more control and more knowledge customers are becoming aware of the privacy issues, the value of their data, and are asking the following question: “If I share this with you, what will you give me in return, and how do I know you won’t use it against me?“
We’ll discuss some of the answers to this questions and I will present several successful and unsuccessful real-world use-cases in this session.
Davor Aničić, STYRIA MEDIA

WNX3 AI in Classifieds and News Media (Styria)

14:50 – 15:20 | HUB385 Academy Room C
Styria Data Science department ( was established in 2015 delivers state-of-the-art machine (deep) learning R&D to Styria’s news and classifieds brands. Utilizing the latest technology and strong data portfolio, artificial intelligence is incorporated into modern media services and products. Successfully launching 1st in the World Visual Search based on the neural network within classifieds industry awarded by NVIDIA and several other AI services for the media industry. Main research domains are Computer Vision and Natural Language Processing.

Coffee Break

15:20 – 15:40 | HUB385

RLS4 Machine Learning Time Series Forecasting

15:40 – 16:10 | HUB385 Academy Room A
Machine learning time series forecasting can have huge impact on companies that sells products, like retail stores, restaurant chains, etc. We will present the case of restaurant chain, where we build a model that forecasts sales of every store in that chain, so the manager can optimize food stocks, so that it minimizes surplus in stock, which in food industry often goes to waste. (Retail Case)

CAL4 IAFS Project: Fraud Detection System

15:40 – 16:10 | HUB385 Academy Room B
Faculty of Electrical Engineering and Computing is developing a IAFS project in cooperation with Multicom, a company from Zagreb. The goal is to develop a generic system for real-time fraud detection. Open source technologies are at the core of the system: Spark, Kafka and Hadoop. Potential fraud events are automatically flagged and blocked to protect users as well as service providers. The ability to adapt to new fraud patterns is very important and data flow is visualized in real-time in the control center. (Telecommunications Case)

WNX4 Assets Digitalization: Blockchain based services Deep Dive

15:40 – 16:10 | HUB385 Academy Room C

RLS5 Bridging the online-offline data gap in digital advertising

16:10 – 16:40 | HUB385 Academy Room A
How to connect online and offline data to achieve uplifts in digital advertising through better targeting. This would cover some interesting technical aspects of how to connect the two worlds, some general data driven marketing information and end results from actually using this connected data in customer cases.
Valentina Đorđević, Things Solver

CAL5 Anomaly Detection in Data

16:10 – 16:40 | HUB385 Academy Room B
This presentation will cover the types of anomalies often met in the data, and comparative analysis of two different techniques that could be used for their detection – Autoencoders and Isolation Forest. After a short introduction to theoretical concepts of these techniques, as well as their pros and cons, the results of their application to data from a telecommunication network will be presented and analysed. (Telecommunications Case)
Grozdana Marić, SAS

WNX5 How to combat increasing payments fraud in a real-time world?

16:10 – 16:40 | HUB385 Academy Room C
Payments fraud is among the fastest-growing and best-known means of generating criminal profit. It has evolved from being committed by casual fraudsters to being committed by organized crime and fraud rings that use sophisticated methods.The sophistication of their tactics makes detecting fraud difficult, especially as the volume of bank transactions grows by about 10% every year. The presentation will provide the application of new trends in fraud detection, such as machine learning and artificial intelligence in the real-time mode and how to use these innovative new tools.

Closing Keynote: Data Drives Product

16:40 – 17:10 | HUB385 Academy
The talk will guide you through how to use data to drive your product and also highlight the ways to identify what is most appealing to consumers and how to optimise your website for large scale conversion.Personalisation helps products deliver a more tailored user experience by presenting relevant content to the users at the right time, there by increasing customer satisfaction, repeat visits and conversion.
Furthermore, the speaker will also discuss the methodology and application of experimentation and how this used to make data-driven product decisions at

Networking and Sharing

17:10 – 00:30 | HUB385 Academy

Day 2, 18.05. Friday

Workshop Day

Over here, you participate in one or more great workshops delivered by individual speakers

Workshop Registration (Kind-Of)

08:30 – 8:50 HUB385
Given that you already registered a day before, no big deal. If someone else is attending on your name, no problem – just let us know!

Welcome to the Workshops!

08:50 – 9:00 HUB385
Welcome to the Data Science Economy Workshops!

Keynote: Picking the Right Tool for the Data Science

09:00 – 09:40 | HUB385
There are many great tools for doing data science. The trouble is there are a lot of great tools for doing data science! This session gives you a framework to help you decide on the right tool to fit your needs so that you can get started sooner. We will use the example of Microsoft Data Science tools to explain you the framework outcomes!

AIOps: How Machine Learning is Transforming IT Operations

09:40 – 10:10 | HUB385
IT Operations (IT Ops) are in for a major change over the next few years. This change is driven by frustration with traditional IT management techniques, methods that enterprise IT Ops teams see as unable to cope with digital business transformation. AIOps stands for Artificial Intelligence for IT Operations. It refers to multi-layered technology platforms that automate and enhance IT operations by using analytics and machine learning to analyze big data collected from various IT operations tools and devices, in order to automatically spot and react to issues in real time.
This talk will focus on the main challenges in IT Operations and how machine learning can help addressing them. We’ll cover the elements of AIOps stack and demonstrate practical applications.

Making OLTP Predictions on CRM/ERP Data in near Real Time

10:10 – 10:40 | HUB385
Making predictions in online transaction processing (OLTP) database application has been long-time driver for many enterprise environment. With CRM/ERP demo on customer data, we will explore the architecture behind, a usefull business case and how to handle predictive models, ingesting and storing data in in-memory columnstore and popular machine learning language, that every data scientist is fond of.

Coffee Break

10:40 – 11:00 | HUB385

WKS1 Deep Dive into Deep Learning (Full Day)

09:00 – 17:00 | HUB385 Academy Room A
As the computer systems are getting more complicated, testers and people responsible for testing have to manage testing efforts that are getting bigger and bigger and changing often. Traditional testing on its own is not enough anymore to keep up with the development. Exploratory testing makes it possible to utilize testing experience and to carry out all the activities more efficient, and thus enables spending the time saved in documentation in a more sensible way. During this course you learn about exploratory testing in an interactive way with practical exercises. Interactive exercises improve the challenging thinking skills of exploratory testing. You also learn about the matters affecting the choices of how to conduct testing and about managing exploratory testing.
Note that this is a full day workshop, with 20 people limit and separate paid enrollment (500kn)

WKS2 Implementation overview: Forecasting natural gas consumption using predictive modelling

11:00 – 12:30 | HUB385 Academy Room B
Data Science has over the years become integral part of every organization. Asking right questions and seeking answers in datasets is a virtue that can be mastered. Workshop will be based on previously presented case study on Forecasting natural gas consumption using predictive modelling demonstrating key methodological and technological aspects. Workshop is a demonstration of one of many possible approaches to gather, analyze, process and visualize data. Two main tracks will be presented:
1. Preparing Data: Data Sources/Database, Collecting data (SQL, R), Analyze Data (SQL, Tableau, R)
2. Predictive Analytics: Predictive Modelling (R), Analyzing Results (R), Validation and Visualization (Tableau)

WKS3 Using Microsoft ML Server for Data Science

11:00 – 12:30 | HUB385 Academy Room C
Microsoft’s ML Server can help you overcome in-memory restrictions of R and allow you to build and deploy models that can be used natively in SQL or accessible via a web service. This demo-packed session will take you through the end to end data science process using Microsoft ML Server so that you can see if it’s right for you.g to our database, wrangling our data, to presenting good looking charts back to the user.

WKS4 Data Warehouse & Big Data: What has Changed?

11:00 – 12:30 | HUB385 Academy Room D
In the past few yearS, one of the biggest questions in the world of analytics is whether the rise of Big Data implies the dawn of Data Warehouses. Even with the development of Hadoop 2.0, the usage of Data Warehouses didn’t drop – on the opposite, having Big Data technologies by their side gave a possibility of augmenting the scope of analysed data. New concepts of hybrid warehouses built on the top of Data Lake architectures have risen and are currently implemented by major vendors. Whether the analytics system is built from scratch or is evolving from the Data Warehouse, there are several approaches to the design and implementation, each with their benefits.
Various tools that enable the interoperabililty of structured and unstructured data on all levels (integration and analytical) have been developed. The workshop will be based on the use cases for different challenged in the analytical system using the tools like Apache Impala.

WKS5 SAS Viya: Built for innovative results from analytics

11:00 – 12:30 | HUB385 Academy Room E
SAS Viya extends the SAS Platform to enable everyone: data scientists, business analysts, developers and executives alike – to collaborate and realize innovative results faster. SAS Viya is a cloud-enabled, in-memory analytics engine that delivers everything you need for quick, accurate and consistent results – every time. Elastic, scalable and fault-tolerant processing addresses the complex analytical challenges of today and effortlessly scales to meet your future needs. Learn what SAS Viya can do for your business and your research.

WKS6 Cost, latency, and interoperability are the three dimension of IoT or IIoT

12:30 – 14:00 | HUB385 Academy Room B
Well will examine the story of missing data, the value of IoT, and a number of standards that are open and proprietary, and how global benefit is achieved as IoT becomes more prevalent, through open data, supporting autonomous ‘things’ communicating with other ‘things.’ While open standards will be a point of view I believe critical, what are our options and risk as we build enterprise solutions. API integration, dataStores in containers with microservices, smart gateways, industry/proprietary standards, and even confidence intervals maybe the tools of choice! Break out your dusty copies of Dr Deming’s guide to statistical process control and the PLC workbench guide, because we are going to discuss factories, and the merging of OT and IT.
Lastly, what would an agenda for a data science education look like? Should data scientists or engineers be licensed, certified? What might the curriculum look like, or entail. We will examine what could be topics for aspiring or practicing data workers today.

WKS7 Analysis of the reader behaviour using topic modelling & Hierarchical Fine-Grained Categorization Based on Image and Text

12:30 – 14:00 | HUB385 Academy Room C
In Styria Data Science department ( established in 2015, we work on delivering state-of-the-art machine (deep) learning R&D to Styria’s news and classifieds brands. Utilizing the latest technology and strong data portfolio, artificial intelligence is incorporated into modern media services and products.
The first topic in the presentation will showcase Analysis of the reader behavior using topic modeling implemented on news media industry. The other topic will present hierarchical Fine-Grained Categorization Based on Image and Text implemented on classifieds.

WKS8 Dependency estimation in large time series datasets with applications in financial industry

12:30 – 14:00 | HUB385 Academy Room D
In the era of accelerated adoption of new technologies and the growing amount of available data, high-dimensional time series are becoming increasingly present and important in many industrial and scientific applications. By measuring dependencies in large time series datasets, valuable insight can be gained into the underlying structures of real world complex systems. This talk will address statistical and computational methods for estimation of high-dimensional correlation structures and their applications in the financial industry, with a focus on portfolio optimization and risk management scenarios.