This week, Manthan is delighted to interview a Forbes Top Influencer in Big Data, Kristen Nicole! Having published her first book, The Twitter Survival Guide, Kristen is now working on her second book on predictive analytics.
So naturally, we sought Kristen’s opinion on what’s coming up in predictive technologies, big data and analytics.
MANTHAN: What are the new and unconventional areas where predictive technologies are being applied today?
KRISTEN: Seeing a lot of dynamic applications for actionizing predictive data analytics in some fun and unique ways. One example is dynamic ticket pricing for the NFL, another is dynamic traffic speeds on a smart highway in the United Kingdom.
Education is another interesting market for use cases. TeacherMatch is applying analytics to predict the right fit of teachers with schools, while Nottingham Trent University is using web-based homework systems to analyze students and customize their curriculum. Similarly, GetWaggle’s online platform provides grade school teachers with student data for predicting learning trajectories for the purpose of customizing education programs for entire classes and individual students.
MANTHAN: In order to deploy predictive technologies, what technologies do consumer-oriented businesses need to consider?
KRISTEN: Sentiment analysis and data mining tools. Hadoop has become a regular technology for managing the storage of large data sets, and Apache Spark is emerging as a viable tool for parsing through that data in an enterprise setting. When gathering consumer data, including mobile app data, home device data, smart car data and social media data, businesses need powerful platforms to parse through it all.
Arrayent is another platform emerging as a market leader, helping businesses orient their consumer-driven sensor and IoT data for business intelligence and predictive services. Leveraging the cloud, Arrayent has also incorporated security and back-end code to simplify the process of sourcing and analyzing data and maintaining consumer device software versions.
MANTHAN: Is cloud computing and cloud analytics helping extend the reach of predictive technologies to more organizations? How so?
KRISTEN: Yes. The ability to work in the cloud is helping to democratize access to data analytics tools and work on a myriad of devices. Cloud is lowering the barrier for the skill sets needed to run data analytics. Automation in analytics software is enabling natural language queries and even anticipating some of the correlations an individual or organization may be interested in when analyzing data.
Such automation also makes predictive technology more affordable, and accessible by more organizations.
MANTHAN: We are seeing a blending of predictive with prescriptive approach to generate better result orientation. What are your thoughts on the best approach to drive business outcomes from such technologies?
KRISTEN: Just as hybrid cloud technology is offering the perks of off- and on-premise data management for the enterprise, a combination of predictive and prescriptive techniques can drive more insightful results. As prescriptive methods are still debatable as a business tactic, I anticipate these methods will continue to evolve with time and experience, enabling more automation with the help of machine learning and human input.
An ongoing challenge for a hybrid approach to predictive and prescriptive methods is data sourcing and synthesization. As more software emerges and becomes automated, more data sources can be incorporated into a hybrid approach for not only analyzing data but actionizing it as well.
MANTHAN: What do you see as the most common roadblocks to successful business implementation of predictive technology solutions?
KRISTEN: Lack of knowledge or an unsure starting point in regards to the software, technology and services that can be applied to their specific data challenges. This can be addressed through third party services that offer consultation on what aspects of the business can most benefit from predictive analytics. Many companies also seek to leverage predictive technology for cost-savings. This is often applied to IT automation where sensors can reduce security risks in the data center, smart building solutions where sensors can automate lighting and temperature to save on utility costs, and marketing and customer service where analytics can lead to more cost-effective solutions with minimal additional resources.
MANTHAN: Do you foresee a time when everyone will use predictive technologies? What will that be like?
KRISTEN: We all use our imaginations – predictive technology is an empirical extension of this natural human behavior. I expect predictive technologies will become services all the way down to the consumer level through fitness, healthcare, gaming and education applications, incentivizing certain consumer actions through rewards-based systems and supporting a consumer’s desire towards self-improvement.
Similar to rewards programs with credit cards, airlines and travel agencies, analytics-based services will become standard in mainstream commerce in an effort to not only retain customers but continually glean information from their actions in a consensual but passive manner.
MANTHAN: With smart machines there is a resurgence of applications that are based on unsupervised machine learning algorithms. What are your thoughts on the adoption of automated decision making processes that use predictive algorithms based on unsupervised machine learning?
KRISTEN: We’ve come a long way since rule-based algorithms, and the merging of artificial intelligence with intelligence augmentation is creating a new world for automated technologies. There are several instances in which the adoption of automated decision-making demonstrates a realistic economic benefit, and as these systems improve, the more they will learn from ongoing, real-time analytics.
I don’t think the human factor will be removed anytime soon, primarily due to the security vulnerabilities in a wholly automated system. As machine-life and human-life become more integrated, the more these systems are at risk for security breaches. Until security reaches a more reliable level of preventative capability, humans will have trouble wholly trusting unsupervised systems for decision making.
Kristen Nicole is Senior Editor at SiliconANGLE.com, and has contributed to publications such as TIME Techland and Forbes. Her work has been syndicated across a number of media outlets, including The New York Times, and MSNBC.