Why is data analytics important? Simply put, it’s because knowledge is power, and speaking truth to power through empirical and impartial analysis of raw data is something which enables us to really see the world in the correct light. We are blessed in today’s world to have the analytical tools, the means to collect, access and draw meaningful conclusions from a wider pool of information than at any time in human history. By looking at our behaviour in aggregate data scientists have successfully modelled decision processing in vision, motor control, language, categorization and common-sense reasoning as complex probabilistic models. This has led to advances in machine learning and artificial intelligence in restricted small worlds with few variables and well-defined probabilities over those variables. These models are even able to predict what we will do next or write next with great accuracy (e.g intuitive text or speech recognition with Alexa, Siri or Google Talk).
In this article, we examine some of the latest trends in big data analytics where It, AI and cognitive computing are beginning to overlap with the commercial world.
Prediction #1: In-store IoT appliances will bring real-time analytics to retail marketing
The burgeoning IoT (Internet of Things) era is changing how the “real world” relates to the digital sphere, with web-connected appliances capable of tracking how devices are being used in real-time. These sensors produce vast amounts of data for analysts to mine, from functional data (power use, performance over time, location) to user behaviour metrics.
A recent blog by IoT analytics provider Mnubo focuses on the example of how smart refrigerators in retail spaces are beginning to shift the way marketers influence customer engagement. Smart fridges can use GPS data to prevent theft or misappropriation; flag signs of imminent breakdown or performance reduction; help to forecast sales by monitoring changes to the number of products on each shelf, and even help to create customized experiences for shoppers.
The great challenge of IoT analytics is understanding which data is directly relevant to actual use cases and business goals. Most retailers purchase products on a regular schedule, estimating needs based on past sales trends. Real-time data from appliances make it possible to calculate needs according to consumption: it can forecast future sales and advise on when new orders should be delivered, taking into account the impact of external factors such as promotions and holidays.
It can even provide insights into the impact of shelf positioning on sales and deliver custom coupons and other promotions to customers at the moment they’re standing in front of the shelf, taking advantage of this crucial moment in the purchasing process.
Prediction #2: Dynamic pricing will make buying staples similar to purchasing gasoline
Every driver has experienced the elation and despair of fluctuating gas prices: when you see the price has gone up since you last drove past the station you regret not filling up earlier, but if it’s gone down you pat yourself on the back. Gas prices change in response to daily shifts in the cost of crude oil, but we will be seeing similar dynamism in the price of other products in the near future.
As Business.com notes, Amazon already updates its prices every ten minutes according to analytics-driven algorithms, while Wal-Mart changes its in-store prices 50,000 times a month. As with the refrigerator example above, the time-to-insight gap is rapidly shrinking for marketers, as sophisticated machine learning artificial intelligence (AIs) turn findings from past trends into recommendations which can be applied instantly. Manufacturers will have real-time access to their competitors’ pricing, allowing them to adjust their own to take advantage of supply and demand.
Prediction #3: Data analytics will have a destabilizing effect on the legal services market
High-skill, high-prestige fields like legal services have been long been fairly insulated from market trends. Due to the opaque and complex nature of the practice, customer reviews are of limited value, as most clients lack relevant knowledge of what constitutes good legal service. Meanwhile, formal rankings like those of Chambers and Partners rely heavily on qualitative evidence: essentially, they conduct thousands of interviews with lawyers about the reputations of other lawyers.
There are now a cluster of Big Data firms led by IBM’s ROSS AI seeking to translate information from court records into empirical data points. This information can be used to create win/loss records for individual attorneys; evaluate the impact of judge and lawyer pairings on case outcomes; and develop an understanding of which venues portend the most favourable outcomes for various types of case. Lawyers and firms with strong track records will be able to use their statistics to market themselves, while others with good reputations but weaker numbers will be forced to grapple with this new challenge.
Prediction #4: We will have more insights into healthcare delivery than ever before
Hospitals are routinely near the forefront of technological development, and IoT is no different. If you’ve ever worn a Fitbit, which monitors vital signs and estimates caloric burn, you’ll be familiar with the concept of wearable health devices. Hospitals are taking the concept to another level by incorporating IoT analytics into medical technology.
AIs can assist nurses by alerting them to shifts in patients’ conditions, setting reminders to deliver medications and even warn of potential overdoses. They can also assist doctors by making diagnostic predictions based on patients’ medical histories and testing. It is arguably administrators and policymakers who will derive the greatest benefit from analytics, as it can help them to more efficiently delegate resources and manage facilities.
Prediction #5: Is managing the influx of Computational Propaganda campaigns critical to the survival of liberal democracy
When it came to light the tech firm Cambridge Analytica had extensively mined Facebook user data (with the full cooperation of the social media giant) to influence the results of the 2016 American election, the discourse was split between those expressing shock, and those shocked that others were surprised. As The Guardian noted in a retrospective look at how companies like CA have turned social media metrics into one of the most significant tools in any election campaign’s arsenal, these analytics provide unnervingly-detailed insights into human behaviour: analytics can predict intelligence, sexual orientation, gender and even incidence of psychological trauma with a high degree of accuracy. Since then an initiative by the University of Oxford computational propaganda project further identified political polarization was more widespread than thought and encompassed many different methods
In the next few years, expect to see the discussion ramp up over how much data Facebook and other giants are allowed to collect on their users, who has the right to access it, and how its potential to disseminate fake news and other political propaganda can be circumscribed.