There are ebook analytics startups like Jellybooks that can tell publishers how members of a focus group are using an ebook – which pages are the most popular, how far each test subject got before dropping the book, etc.
But one thing these startups can’t do just yet is tell publishers how a reader feels about a book. Sure, they can survey readers, but that doesn’t always result in truthful answers.
But sentiment analysis might.
According to Wikipedia, sentiment analysis “refers to the use of natural language processing, text analysis, computational linguistics, and biometrics to systematically identify, extract, quantify, and study affective states and subjective information”.
A simpler way to put it would be that sentiment analysis is when someone uses software to identify the emotional subtext in a text, image, or video. For example, Facebook regularly uses sentiment analysis to understand and filter your updates.
The part I want to bring to your attention is how someone might use a camera to watch faces, and then use sentiment analysis software to understand what people are feeling.
A story on the topic crossed my desk today. IHE reports on one such effort to use sentiment analysis in the classroom and track students’ engagement levels.
How’s everyone doing so far? Am I being clear? Anyone confused?
Professors might ask these questions midway through a lecture to get a sense of students’ moods. The scattered answers often aren’t very helpful, if they’re even accurate.
With sentiment analysis software, set for trial use later this semester in a classroom at the University of St. Thomas, in Minnesota, instructors don’t need to ask. Instead, they can glance at their computer screen at a particular point or stretch of time in the session and observe an aggregate of the emotions students are displaying on their faces: happiness, anger, contempt, disgust, fear, neutrality, sadness and surprise.
The project team hopes the software will help instructors tailor their teaching approaches to levels of student interest, and to address areas of concern, confusion and apathy from students. If most students drift into negative emotions midway through the session, an instructor could enliven that section with an active assignment. If half the students are happy and the other half aren’t, the latter group might be getting left behind.
Consider for a moment how it might help publishers and authors understand readers.
The camera on a reader’s smartphone or tablet could track the reader’s face, taking a photo every few seconds. Those photos would then be sent to the publisher’s servers where sentiment analysis software could identify, for example, which sections of a book were the most exciting.
Yes, that is creepy, but it could still be useful for publishers who want to know whether a story hit the right emotional buttons.
So what are the chances that someone might use this?
I asked Jellybooks founder and CEO Andrew Rhomberg what he thought about using this software, and he told me that “I think that would be way too spooky even for a typical test reader,” adding that he didn’t plan to implement this because “there is so much other lower-hanging fruit (in publishing) we can and should solve with technology first.”
That is reassuring to hear, but we should not be complacent on this issue. The best way to stop anyone from using sentiment analysis on readers is to object to its use the first time it is proposed, which is why I brought this to your attention today.
There will come a day where some tech startup will offer a service that can track readers’ emotional responses, and it will be on authors and publishers to say “not with my books”.
What do you think of sentiment analysis? Can you see an upside, or is this pure creepy overreach?