Skip to main content

The Ocean's Big Data Platform

André Karpištšenko (Proekspert)
Data Driven Business
Ballroom CD
Average rating: ****.
(4.22, 9 ratings)
Slides:   1-PDF 

Marinexplore is building a big data platform for accessing, analyzing and managing oceanic, geophysical and atmospheric data. The cloud solution simplifies complex data processes for the users who can focus on collaboratively solving important business decisions and in-depth analyses instead of spending time on data wrangling tasks.

The problem is characteristic to many big data spaces. There are 2+ million sensors deployed in the oceans, spread over thousands of data sources, majority of it public, but also private. These come in a variety of formats and conventions. 80% of the time related to decision-making is spent on finding, connecting, cleaning, creating data products, analyzing and visualizing this data. All before proprietary models are run against selected data using in-house analytics, HPCs, etc. So it is critical to cut time spent on these activities, by streamlining data flows in a smart way.

Marinexplore integrates, analyzes and visualizes diverse data ranging from simulations output to observations done by satellites, radars and sensors in the environment. Considerable data integration effort is spent on ensuring compatibility between different standards, protocols and data models. Associated metadata is preserved next to original data via a flexible automated plugins based collection mechanism. Data is stored in a purpose built solution based on numerous maturing data technologies. Content is made accessible to people via an exploratory visual tool. API allows connecting other data systems and tools (e.g. Python, R, Matlab) directly to the cloud services.

With considerable amount of ocean data being open and sensor technologies rapidly changing, Marinexplore has built a community of ocean professionals. The community participates in developing the capabilities of the platform. As a result it is possible to have better understanding of changing conditions at sea by running analytics on private and public data combined.

Marinexplore as a company was founded in February 2012 in US, with a second office in Estonia. The team has strong expertise in software engineering, oceanology, energy and automation, robotics and product design. Marinexplore is co-chairing the World Ocean Council’s SmartOcean program.

Photo of André Karpištšenko

André Karpištšenko

Advanced Products and Technologies Lead, Proekspert

Co-founder and Chief Scientist of Marinexplore leading engineering and research efforts to create the ocean’s big data platform.

Previously founded the Data Research Team of Skype, applying data analytics, predictive methods and system optimizations to improve the products. At Skype, he built 3 teams, one of which was responsible for initiating and delivering a number of engineering toolset improvements spanning the entire organization.

Prior to Skype, he founded & built up a Quality Assurance consultancy and developed Electronic Ticketing Services of Tallinn. He was also involved in developing the early versions of Eionet systems for European Environmental Agency.