The business of sentiment is permeating our lives
By Miran Varga, Delo, 08.05.2017
Translated into English
Sentiment analysis will completely change the perception of business and consumers, and could have the most beneficial effect in health care.
Humans are sentimental beings, dealing with a multitude of different emotions each day. Numerous companies and their marketing and advertising campaigns target our emotions, trying to entice us to buy a product or service. As it is, people make purchases using their emotions. Using modern technologies, various data can be transformed into emotions. Or in other words: our emotions can be described using data that can be interpreted by various analytical systems.
At the SAS Forum in Milan, experts from the aforementioned company presented the levels of progress in sentiment analysis. The focus on consumers’ emotions is more than obvious in shops, where technology enables the recognition of consumers’ sentiment and mood by using video cameras and facial analysis, and even an individual’s posture. Next, the responsibility for drawing attention is assumed by “active” shelves and advertising material, which, in accordance with the individual’s identified sentiment and interest – or lack thereof – present an appealing offer.
When used for health-care purposes, a lot of data can be gathered by smart bracelets monitoring the patient’s heart rate, skin temperature and pH, condition or movements, as well as other parameters (some can even monitor sleep). Generally, these devices transmit data via Bluetooth to smart phones or other devices for data collection and processing, where sentiment analysis then performs its “magic”.
Mobile applications reveal to retailers everything about the consumer
In this age of smart mobile devices, this is made that much easier. If a consumer uses the retailer’s mobile app and/or makes purchases through the retailer’s online store, the retailer learns practically everything about the individual and their shopping habits and preferences. Thus, the retailer can send a personalised, attractive offer to the individual as soon as the retailer’s system detects the user in the vicinity of the shop. This scenario, which may seem like it belongs in a sci-fi movie, is already being tested by numerous progressive retailers.
Especially progressive are some Asian and American apparel retailers, who have begun implementing so-called smart mirrors in their stores. Such mirrors allow customers in the store to quickly see how they would look with new clothing, a new hairstyle, etc. Of course, all of this makes the whole user experience much better, much more attractive, resulting in the consumer making the purchase and returning to the store. This is very important in the digital and online world, where loyalty is a very rare virtue.
Sentiment in education, banking, health care ...
The use of sentiment analysis is expanding to fields where we might not even expect it. Lecturers can use sentiment analysis to monitor the mood of their audience and appropriately adjust their presentations. Sentiment analysis tests that monitor the attention of listeners or viewers in online course programmes lead to excellent results, even more so in active participants, i.e. those that make contact with the lecturer.
Adapting courses to online students or listeners enabled adjustments to be made to the learning curve to adapt to specific online audiences and gain better results (e.g. a higher percentage of successfully completed exams).
A very interesting example of sentiment analysis use is also the use of video cameras at ATMs. Some banks have already implemented a system that recognises “stress” activities in the ATM’s field of view and triggers an alarm.
Sentiment analysis will be very useful in health care as well
It seems that sentiment analysis will do the most good in health care. Here, the importance of patients’ emotions is severely underrated, since most medical devices and medications treat only the physical effects of a disease, while only psychologists in exceptional circumstances address patients’ emotional responses to a disease and treatment.
The Canisius Institute for Autism Research found that machine-based detection of patients’ emotions and appropriately adjusted responses can lead to significant improvements of this condition, especially with children. The Multiple Sclerosis Association of America made similar findings. A high level of irritability and frequent mood swings with such patients lead to great distress and anguish; however, the use of technology that can recognise anxiety and irritation in patients makes it easier to maintain their well-being.
How does sentiment analysis work?
In technical terms, sentiment analysis is very similar to data-flow analysis, as it analyses a vast amount of data from various sensors in real time, presenting results that can be used by commercial users, advertisers, physicians or other automated systems.
When used for health-care purposes, a lot of data can be gathered by smart bracelets monitoring the patient’s heart rate, skin temperature and pH, condition or movements, as well as other parameters (some can even monitor sleep). Generally, these devices transmit data via Bluetooth to smart phones or other devices for data collection and processing, where sentiment analysis then performs its “magic”, comparing current samples to pre-set health values, and recognising potential deviations and anomalies. If recognised, these can be marked as events that are hazardous to the patient’s health, with notifications being sent to both the patient and the treating physicians.