Accelerate the Transformation

Everyone talks about how machine learning will transform business forever and generate massive outcomes. However, it’s surprisingly simple to draw completely wrong conclusions from statistical models, and “correlation does not imply causation” is just the tip of the iceberg. The trend of the democratization of data science further increases the risk for applying models in a wrong way. In this talk, we will discuss capital mistakes as well as small errors that add up to completely ruin the potential positive impact of many data science projects. Attending this talk will hopefully help you to avoid many of those mistakes.

Blockchain is a new technology for connecting partners that not completely trust each other. In Enterprise systems every partner in a blockchain needs to connect his individual systems of records, create complete new user interfaces or both. Mostly there is also additional data that needs to be stored “off the chain“. Backend as a service provides possibilites to store data, create APIs, Frontend SDKs with analytics and even connect existing Backends like ERPs, CRMs or databases with out of the box connectors. Both technologies combined can reduce effort in Blockchain projects. Examples will be shown in the presentation.

“Blockchain-based Distributed Ledger Technologies (DLT) are rapidly transforming established B2B and B2C cooperation patterns; innovative solutions range from ledger-based non-repudiable communication to radical disintermediation scenarios.

There are many exciting and disruptive new possibilities at the intersection of DLT and data-driven cooperation, with applications from IoT and cyber-physical systems through healthcare to supply chain management. However, at the same time, real life applications have to fulfill a number of requirements beyond functionality - e.g. data visibility, privacy, process orchestration and performance - that make solution development a much more nuanced task than “smart contract programming“ in itself.

The presentation gives a broad overview of these novel application patterns, and demonstrates how Blockchain technologies of the Linux Foundation Hyperledger project can be used as a framework to realize data-driven cooperation solutions that fulfill key enterprise and industrial requirements.“

Big Data vendors often say: “Save all of your logs, or any kind of information as raw data. Storage is endless and you never know what data could come handy in the future, or in what format you’ll need it.” However, when you want vendor support and use quality hardware, Big Data suddenly becomes expensive. It feels even more expensive because you won’t ever use the majority of the data. Using syslog-ng for logging can optimize your Big Data costs in multiple ways:
- Parse and enrich log messages (or any kind of collected data) to help message filtering. Discarding useless messages can significantly reduce the amount of data to be stored.
- Convert messages to the format that the processing applications use, or save them in a structured format that is easy to process, for example JSON. Syslog-ng can do this in real time on the clients, saving you processing time in the Big Data server farm.
- Collect logs and forward them directly to your Big Data destination (for example Elastic, Apache Kafka, Hadoop), without having to install additional applications, simplifying your infrastructure.
- The processing capabilities of syslog-ng are much better than most Big Data client applications, it can process more messages using less resources.
- Processing, reformatting, and filtering messages saves on the costs storage, maintenance, and processing power of your Big Data cluster.

If you’re not into eSports, some of the stats surrounding the rapidly growing industry can be surprising. Electronic video gaming has extended from being a hobby into a serious sport and business. Every day millions and millions play these games or watch them just as spectators, and during these activities produce a significant amount of data, which can be easily obtainable. The data is stored in replay files and made accessible online, providing us the opportunity not only to visualize and analyses them, but also to dive into the details at a level that is unprecedented in other sports. In my speech I will present our solution for an after-game analytics and how commentators can make the most out of it.

When we started originally our first application to create a SaaS solution for energy management we didn’t anticipated for such quick ramp up in demand. The application was aimed for large, multinational companies where it can bring significant saving and fast return on investment. We did a handful of implementation in the last 18 months and have 3 major branches of the same core solution.
As we get more traction even with mid-size companies, where our value-added intelligent analytics can generate significant returns while reducing manual work, we started to rethink our architecture. Our current cloud-based, single-tenant architecture runs on Amazon, leveraging Kubernetes, Angular and handful of current technology. Our aim is to integrate the branches into one multi-tenant application, where we can provide scalable, tailor-made services more cost effectively while lowering the operational requirements of the system. The presentation will go into the whys and hows of this journey.

The presentation gives a rational abstract of chained applications like:
- Parameter settings for automated alert-generation systems
- Possibilities of suspicion generating based on data assets about energy uses
- Development of a controlling/monitoring system for basic stations in the telecommunication
- Possibilities of automated outlier detection in time series in case of energy consumptions

Most of the employees of any organization in the world have only a few data points about how they work every day - the time they start and finish work, plus the task(s) they work on during the day. World-class employees want to know more, they look at how they spend their time, they need deep insights so they can make data-driven decisions. Come and see how Crossover scales up the organization in a fast-paced environment, by creating world-class teams and taking productivity to the next level.

Modern machine learning approaches brought new opportunities in various fields where sufficient training data is available and these methods enabled us to solve extremely complex tasks, but human learning still outperforms algorithmic approaches in various aspects.
We can learn new concepts from a handful of examples, meanwhile when the amount of labeled data is limited -or the generation of a large dataset would be extremely resource-consuming-, Machine Learning approaches tend to reveal their weaknesses.
The aim of this talk is to introduce a couple of techniques which can help increasing the generalization power of neural networks and reduce their strong dependence on extreme amount of data.

Collecting data is the basis of data analytics. If you don't have accurate, automatically collected, real-time data, then data analytics doesn't make much sense. Recently many studies have been published referring to the main challenges and barriers to start the industrial IoT journey. These challenges are related first of all to Connectivity (collecting real-time data directly from the processes), secondly to create a bridge between IT and OT (Operational Technology), and last but not least to Cyber Security. During my presentation, I will introduce some ideas and best practices how to handle real-time data acquisition and how to connect OT systems like PLC, SCADA directly to the Cloud for data analytics.

How NEXOGEN help Waberer's transcend its industry: “The difficulty lies not so much in developing new ideas as in escaping from old ones.“ - John Maynard Keynes
Waberer’s is the European leader in full truckload transportation with an industry leading 92% utilization rate. The company has achieved 18% growth in revenue with a 16.5% growth in profit in 2017. This is a story about a small team with a big dream and their journey to tackle challenges in optimization and automation in the transportation logistics industry with Waberer's. They are called NEXOGEN, and their dream is to help shape a world where all companies are built around smart algorithms to automate and optimize their operations. Turning data into actions and visions into reality.

During the presentation we will follow through how Magyar Telekom created the Data Lake. We will cover the challenges faced and overcome during implementation and operation. We will discuss how a DWH and a Data Lake can co-exist, and how Data Science activity can augment traditional development lifecycles in a truly enterprise production Big Data environment.