Data Science: What to Expect in 2019
You can expect to see a lot of change and a lot of AI implementation in the field of data science in 2019.
Data science is rapidly changing. New advances in AI and machine learning mean that data can be applied in brand new ways, and in unprecedented modeling systems, to do much more than was possible just a few years ago. The cloud is also ushering in a new era of data science by making software more portable and versatile.
Techopedia asked the experts what we might see in the year ahead. Here’s some of what’s likely to come our way in 2019.
“Demand for smart analytical applications will redefine enterprise data practices: Enterprises are in a race to become data-powered businesses, yet only a small fraction of the value of advanced analytics has been unlocked. In 2019, there will be high demand for new innovations around smart analytical applications that are driven by real-time interactions, embedded analytics and AI. …
“The rise of the data engineer brings AI to the forefront within the enterprise: Last year was the year of the data scientist. Enterprises focused heavily on hiring and empowering data scientists to create advanced analytics and machine learning models. 2019 is the year of the data engineer. Data engineers … specialize in translating the work of data scientists into hardened, data-driven software solutions for the business. This involves creating in-depth AI development, testing, devops and auditing processes that enable a company to incorporate AI and data pipelines at scale across the enterprise.
“Human and machine learning form [a] symbiotic relationship to drive real-time business decisions: In 2019, the world of AI and analytics will need to converge in order to drive more meaningful business decisions. This will require a common approach for combining historical batch analytics, streaming analytics, location intelligence, graph analytics, and artificial intelligence in a single platform for complex analysis. The end result is a new model for combining ad-hoc analysis and machine learning to provide better insights faster than ever before.”
— Nima Negahban, CTO and co-founder, Kinetica
“Developers [will] learn they need a data scientist friend.
“Developers aren’t going to become data scientists – one writes code, one thinks in math and models. But devs will increasingly need to understand data science methodologies, and to integrate data science models into their workflow. Data makes software more intelligent, giving it the ability to predict outcomes or anticipate user needs through machine learning. So developers increasingly need a new level of partnership with data scientists to make great work happen. Developers can expose the data scientists’ models via APIs, and embed them in domain-specific apps to really drive change.
“Consider a retailer trying to intelligently decide which brick and mortar store to fulfill e-commerce orders from. A data scientist can create the model that calculates the optimal store from which to ship, so that the company ships the sweater that was likely to sit on the shelf at a store in a warm location, rather than the one likely to get bought by an in-store shopper in frigid areas. A developer could pull that kind of intelligence into a fulfillment app, and put it in employees’ hands to make the right decision.”
— Siddhartha Agarwal, Vice President, Product Management & Strategy, Oracle Cloud Platform
“In 2019, artificial intelligence (AI) and machine learning (ML) will nearly reach its full potential by connecting and processing data faster over a global distribution of edge computing platforms. AI and ML insights have always been available, but possibly leveraged a bit slower than needed over cloud platforms or traditional data centers. Now we can move the compute and storage capabilities closer to where data is retrieved and processed, enabling companies, organizations and government agencies to make wiser and faster decisions. We’re already seeing this in the way airlines build and service airplanes, government defense agencies respond to hackers and how personal assistants make recommendations for future online purchases. This year, thanks to AI and ML, someone will finally know if that special someone really wants a fruitcake or power washer.”
— Alan Conboy, office of the CTO, Scale Computing
“2019 seems as if it will be the year of analytics, machine learning and AI. These tools are already available, though their take-up has often been delayed by a failure to match these new capabilities with appropriate new workflows and SOC practices. Next year should see some of the pretenders – those claiming to use these techniques, but actually using the last generation’s correlation and alert techniques in disguise – fall away, allowing the real innovators in this field to begin to dominate. This is likely to lead to some acquisitions, as the large incumbents, who have struggled to develop this technology, seek to buy it instead. 2019 is the year to invest in machine learning security startups demonstrating real capabilities.”
— Stephen Gailey, solutions architect, Exabeam
“As AI and ML become mainstream, a new breed of security data scientists will emerge in 2019: AI and ML techniques are data dependent. Preparing, processing, and interpreting data require data scientists to be polymath. They need to know computer science, data science, and above all, need to have domain expertise to be able to tell bad data from good data and bad results from good results. What we have already begun seeing is the need for security experts who understand data science and computer science to be able to first make sense of the security data available to us today. Once this data is prepared, processed and interpreted, it can then be used by AI and ML techniques to automate security in real time.”
— Setu Kulkarni, Vice President of corporate strategy, WhiteHat
“In software development, the big story in 2019 will be machine learning and AI. In the coming year, the quality of software will be as much about what machine learning and AI can accomplish as anything else. In the past, delivery processes have been designed to be lean and reduce or eliminate waste but to me, that’s an outdated, glass-half-empty way of viewing the process. This year, if we want to fully leverage these two technologies, we need to understand that the opposite of waste is value and take the glass-half-full view that becoming more efficient means increasing value, rather than reducing waste.
“Once that viewpoint becomes ingrained in our M.O., we’ll be able to set our sights on getting better through continuous improvement, being quicker to react and anticipating customers’ needs. As we further integrate and take advantage of machine learning and AI, however, we’ll realize that improving value requires predictive analytics. Predictive analytics allows simulations of the delivery pipeline based on parameters and options available, so you don’t have to ‘thrash’ the organization to find the path to improvement. You’ll be able to improve virtually, learn lessons through simulations and, when ready, implement new releases that you can be convinced will work.
“Progressive organizations, in 2019, will be proactive through simulation. If they can simulate improvements to the pipeline, they will continuously improve faster.”
— Bob Davis, CMO, Plutora
“In 2019, look for data teams to become more sophisticated as their field matures, evolving to work with bigger data sets and integrating new techniques into their workflow. Advanced languages like R and Python have become a more critical part of everyday analysis and should be a centerpiece in the ‘day zero’ strategy when building any technology stack.”
— Harry Glaser, CEO, Periscope Data
“I expect to see widespread adoption of a method I recently wrote about, Mixed Formal Learning. It enables companies to create AI systems with outstanding accuracy and zero or tiny amounts of training data. From a raw materials standpoint, the sheer mountain of data exhaust required to train and test AI solutions prevents many companies from entering the AI race.
“At least two companies, Google and Glynt.ai, have demonstrated spectacular results using Mixed Formal Learning. Glynt.ai uses this method to extract data from unstructured documents with fewer than 10 training examples. The result is accuracy of about 98%: better than a team of two data entry clerks. Previous implementations would be proud if they needed 1,000 examples to do the same task with accuracy of 95%.”
“If I have to predict what will be big in 2019, I would say AI – I think we'll see much better virtual assistants and better chatbots. The thing with AI is that we, the consumers, are not really aware of the rate of growth of the technology and its applications, because it’s working behind the scenes.
“Also, blockchain is big, and it’s getting bigger. For now, it doesn’t have applications in data science, but I won’t be surprised if in 2019 it starts to have [applications]. For sure, all this decentralized storage could be harnessed to serve big data.”
— Vania Nikolova, Ph.D. in Mathematical Analysis, Head of Data Analytics at RunRepeat.com
“Recent massive investments in data science should significantly change the social media landscape during the next few years. As a social listening vendor, we see an emerging interest in AI-powered image recognition technologies, introduced by major SML vendors. These are already being used by some brands which are early adopters and innovators and are ready for global demand.
“This technology brings a whole new level of marketing insights to consumer brands and agencies. It helps them to understand better their consumers’ tastes – even though the brand is not mentioned explicitly in their social media posts. It is an excellent way for marketers and social media specialists to learn more about product consumption situations, and uncover valuable consumer insights.”
— Alexandr Sirach, co-founder at YouScan (AI-powered social media listening platform)
“In 2019, we see the demand for data science pipeline platforms increasing dramatically. We often compare data science to the software development boom, where GitHub and other software development platforms drastically impacted the field of development. We see data science platforms escalating and leveraging the data science field.”
— Yochay Ettun, CEO and Co-founder of cnvrg.io and expert in Data Science and Machine Learning
“We predict that 2019 will be the year data science goes mainstream. Many of our clients and partners have started to kick the tires on deeper data science initiatives in 2018, and we feel that the momentum is growing quickly to integrate data science into organization-wide decision-making and policy by the end of 2019.”
— Sam Underwood, VP of Business Strategy with Futurety, an Ohio-based data analytics and marketing agency