21 Aug 2018
Machine Learning Set to Optimise Big Data Analytics Processing Time
- Photo: The Cloud: Set to be at the heart of Big Data analytics and post-analysis strategic implementation.
- Photo: Big Data: Digitally detecting dementia biomarkers.
- Photo: Handwritten legacy data: A processing challenge.
- Photo: Big Data World 2018: Every company is now in the business of data application.
With Big Data seen as an essential tool in many business environments, smart-processing systems, coupled with enhanced legacy data migration capabilities, are set to unleash the full potential of commercial information streams.
Big Data is delivering real benefits, especially to marketers and medics, at least according to many of the keynote speakers addressing the delegates at this year's Big Data World event. In the Big Data sector, interest was particularly high with regard to the role machine learning is now playing in automating data analysis, with it now being far easier to glean actionable insights from large data sets without the labour and time-intensive number-crunching that was once obligatory. Despite this somewhat huge leap forward, however, there remained, for many, the immense challenge of effectively migrating legacy paper records into a processable electronic format, a transition that still seems to intimidate a number of business owners.
This year, the event – co-located with Cloud Expo Europe, Cloud Security Expo, Smart IoT and Data Centre World – attracted an array of international delegates, all assembling in London on a mission to share insights, explore new technology and catch up with all the latest IoT innovations. As usual, much of the event's educational component came courtesy of the show's extensive seminar / workshop programme, which this year boasted more than 100 high-level speakers from the world's various IT pinnacles.
Within the dedicated Big Data World division, it was the healthcare industry that dominated proceeding with its twin focus on improving diagnoses via better-harnessed Big Data and effectively managing the digital migration of the huge existing volume of paper-based patient information. Outlining the progress made by his own department, Professor Sebastian Ourselin, Head of University College London's Translational Imaging Group, talked delegates through just how the institution was now processing MRI-derived Big Data. Although still at a relatively early stage, he maintained that considerable progress had already been made with regard to identifying the biomarkers that correlate with heightened susceptibility to such conditions as dementia or prostate cancer.
In order to bolster the accuracy of such predictive diagnoses, however, a way must be found to cost-effectively digitalise the mountain of patient data amassed in more analogue times, with the UK healthcare sector working to a provisional 2020 deadline for completing this massive process. A presentation by Omar Divina, Sales Vice-president for HyperScience, a New York-based AI developer, however, offered some hope for those tasked with delivering this vital transition.
Defining both the problem and the solution, he said: "Deciphering patient records, which to the lay person often seem almost illegible, is a unique challenge. In a bid to tackle this, we have developed a document scanner that has the ability to 'learn' a person's handwriting, enabling it to capture up to 90% of all handwritten material."
Even when all the data exists in a digital format, analysing information flows from a variety of sources can still be far from straightforward. Looking to tackle this particular challenge within the sales and marketing sector was SplashBI, a Georgia-based specialist in business analytics.
With a system than can apparently harvest the relevant trends and indices from across a number of platforms, including Google Adwords, Marketo and Hubspot, the company claims to offer Big Data insights at an unprecedented speed. Summarising its proposition, Chief Marketing Officer Marc Ramos said: "We can help integrate, correlate and analyse all your data sources via one platform, presenting real-time intelligence as required."
Cutting the information-lag while optimising relevance was also part of Qubole's offer, with the California-based company keen to demonstrate just how its cloud-native Big Data activation platform could rapidly deliver meaningful results. Outlining the system's USP, Solution Architect Tejeswi Mallkarjuna said: "Data projects can take longer than expected, eating-up man-hours and then not delivering any actionable results. Our platform, however, enables customers to self-manage and self-optimise their data in order to deliver actionable insights."
Another US-based company with a focus on maximising the benefits of Big Data in the marketing arena was North Carolina's Core Compete. Defining its particular role, Sunil Adlakha, Head of the company's newly launched London office, said: "Despite having been in business since 2012, we still see ourselves as a start-up, partly because we are continuously developing our offering for clients as we work with them to maximise the value of their data. We pride ourselves on creating valuable business insights for our customers, with our analysis making a real contribution to their sales and marketing activities."
Another business in expansionist mode was California-based SnapLogic, the company behind the Dataflow scalable data integration platform. Providing a snapshot of the company's current positioning, Vice-president Craig Stewart said: "We are in the process of finalising a round of multi-million-dollar venture-capital funding. We see this as a prerequisite for the further development of the Dataflow platform as we look to enhance its on-demand data-integration capabilities."
Neatly encapsulating the overall business benefits of the emerging generation of Big Data processing technology, Paul Collins, an Account Director with Avora, a London-headquartered business intelligence service provider, said: "Increasingly, companies are coming to realise that business data analytics is the key to maintaining a competitive advantage. In order to be truly effective, managers need to ensure that their team has all the necessary tools to fulfil the role of data scientists when required.
"For our part, our system can consolidate data from more than 300 sources, while our machine-learning intelligence can detect both data anomalies and trends, highlighting them in order to ensure all subsequent business performance decisions can be duly optimised."
Big Data World 2018 took place from 21-22 March at London's Excel Exhibition Centre.
David Wilkinson, Special Correspondent, London