Jens-Peter Seick, Fujitsu
Jens-Peter Seick, Fujitsu
We already have the technology; now we need the uses
The second half of the Summit covered the vendor side of big data and its role in research and development (R&D). Speaker Jens-Peter Seick, Vice-President of Product Management and Development at Fujitsu in Europe, leads R&D and management of the company’s products. Big data and big data-collecting products from Fujitsu can make life easier for end-users, he explained.
He began by explaining that big data must be seen in the scenario of a hyper-connected world. The interconnectedness of devices – the internet of things – creates a huge amount of data every day that has the potential to increase business insights and value for end-users and for businesses; there are already technologies for analysing it, but what is needed is uses for it.
Big data was first used by the retail industry. “Here it’s a competitive threat not to use it for predicting end-buyer behaviour,” said Seick.
Newer examples include airlines that collect aircraft performance data at the rate of two terabytes (1012 bytes) of data per hour of flying; the data is integrated into a system for analysis that aids predictive maintenance, and is used by flying crew and ground maintenance crew. Big data is also used for predictive maintenance of offshore wind turbines arrays, because it's expensive to send a maintenance ship and crew to a single defective offshore wind turbine. It’s much more cost-effective to do so as a planned maintenance exercise.
“Data scientists are like unicorns – they only exist in fairy tales”
Businesses don't always have a data scientist on board, nor do they have a precise idea what can be done with big data, or how it can be combined with other data sources. But it’s a threat to the competitive advantage if businesses don’t use it, said Seick. Customers ask what they can do with the data they already possess, and how it can be combined to create additional value. What process management and architecture does it need to become more useful?
“These data scientist are like unicorns, you only find them in fairy tales!” he commented.
External data can add value
The idea of the business value customers can create with big data is of key importance, and even if the target is still unknown, the process is clear. Existing data from a variety of sources can be collected and cleansed to discard unwanted data. Then it can be analysed – which can also yield predictions. “You can make decisions based on it, but even then you are not done,” he said. There could be several cycles of collecting and refining data until the target is reached, he explained.
The quality of the process needs data scientists who understand how the business and the data are organised – and how external data can add value.
The recent ALS ice bucket challenge is an interesting topic, said Seick. It had raised the profile of the medical condition and had also raised the possibility of collecting patient data “because more data is needed to find treatments. Researchers need higher quality data to solve the problem faster.”
Big data generates answers to business problems
The technology associated with big data is undergoing evolution. “It's not rocket science,” he said. “Storage and collection is well understood, and there are no restrictions on the technical side. But in terms of processes, it's an extension of intelligence. What you’re doing differently is integrating data from different sources. Business intelligence is big data generating answers to business problems.”
Seick gave an example. Fujitsu has used its infrastructure to find the best location in the Baltic Sea or North Sea for a European offshore wind farm. The data source was European mid-sized weather forecasting systems.
More than a million files containing weather data were cleansed and transformed before being used to create a visualisation for those not familiar with that type of detailed knowledge. “Then you can play with the parameters of temperatures, water depth or wind speed,” said Seick. Information about earthquakes from another institute or about shipping from another source can be added to improve the quality of the information.
The final result, said Seick, is “when you have no more ideas and you have reached the highest level of business value. Then you can make your decision. So far, so good!”
Asking the right questions
But in terms of technology and implementation, how can the right question be asked of the big data exercise, and how can we be sure that the data adds value?
Companies preparing the business case for a big data project have used costs for individual products from different vendors, and found their projects were rejected by financial directors because the return on investment (ROI) was not immediately visible. “That doesn't make sense,” said Seick. “Their clients can't fathom what decisions can be made from it and what value can be created from the information that is really in the data. If you can't calculate the ROI you can't get permission to make the investment.” Fujitsu has created the Apache Hadoop integrated system to distribute the analysis of big data over several (or thousands of) servers to meet a real business need.
The smartest company will win
Seick said Fujitsu infrastructure can help to keep ROI under control, and recommended smaller steps: iterative analytics, as opposed to classical business analytics. “Ask the right questions one at a time, and establish value at each step. Do in-house tests, even if they are trial and error at this stage, and you increase your value with each cycle.”
Jens-Peter Seick invited students and researchers to propose projects for Fujitsu’s products and invited them to visit the company’s exhibition stand at the Summit.
Examples of such projects include:
- Fuel price behaviour analysis, which examined and compared daily, weekly and seasonal prices changes among neighbouring German fuel stations to find the best time to buy fuel.
- Investigating the quality of short-term and long-term weather forecasts; this revealed the best service provider.
Big data is a reality and is implemented in many verticals, he concluded. Customers often don’t know what information is contained in their big data, and therefore don't get the benefit of it. New technologies are developing fast, and the infrastructure is easy to understand and quite mature. “Technology is not a hurdle any more, and the smartest company will win,” he said.