There are a plethora of new companies working on mood awareness in technology, especially in the music software world. Apps like Songza take various inputs into account, like time of day or weather, try to guess your mood, and then play context appropriate music. That’s pretty cool. It’s another level of personalization, an answer to an ever growing (perceived? created?) desire for customization. At the very least it’s a differentiator. Nevertheless, it’s a lot of amazing technology set up to just choose a bunch of songs for you, instead of deciding for yourself what you want to listen to. Like a personal digital DJ that’s trying to read your mind.

But what else can we do with mood detection? What about telling you what your mood is? Or how about how your mood is perceived by others?

The Internet is renowned for, among a great many things, the ridiculous flamewars people get into as a result of anonymity and the inability to create or decipher subtlety and nuance from words. People are constantly misinterpreting each other’s words, getting offended, and getting into absurd, useless arguments. Part of this is people’s willingness, even eagerness to get into a written scrap, but the other half is our own difficulty with written communication. It’s tough! We’re not all professional writers, and those that are have years, decades of experience. Miscommunication is inevitable!

What if we could help cut down on that miscommunication from the source? Once you’ve written something, be it a blog post, a blog comment, a social media update, whatever, what if something looked it over automatically and told you what mood it thought you were in, what mood you were trying to express? It could take into account the writing stage, the review stage if any, the context of the writing if present (like a blog comment). It could then present you with a quick report on what emotion you’re writing embodies, what others might get out of it. If it’s exactly what you wanted, great! Hit submit. If it’s way off, you have a chance to correct it before strangers on the Internet start wishing you bodily-harm.

This can serve to not only make us more aware of how we present ourselves, possibly improve our writing in general, but also to develop more empathy toward one another. If we see how easily we can be misunderstood, how quickly our innocent comments become acts of war in another’s eyes, perhaps we’ll be a little more forgiving with our judgements.

On the technical side of this, I imagine the mooding algorithm would take into account:

• Typing speed
• Length and location of pauses
• Number and frequency of mistakes
• Number of corrections to those mistakes
• Number of words per sentence
• Punctuation
• The content, of course
• The context of that content
• Mouse clicks and selections within the textbox
• Time between the last words typed and the submission
• Historical context of the above elements and the resulting analysis

Possibly to a point of excess, it could also take in:

• The time of day and local weather
• The point of the user’s wake/sleep cycle, calculated by Internet activity throughout the day or night
• Lighting conditions in the room (webcam or ambient light sensor in smartphones)
• Ambient noise level and media playing

This would obviously need to be a learning algorithm. It would gather data from each user, and use it all to compare moods in general, and use data specific to a person to calculate their specific needs.

Combined, this gives an idea of not just the written mood, but the mood of the user as well. Are they writing in a fury, misspelling things left and right, hitting submit before their rage settles? Are they carefully choosing each word, plodding along each sentence? Are they eagerly scrawling their note of appreciation with a few too many exclamation marks?

Finally, similar to up- and downvoting or liking systems, there could be a mood system integrated into social media. Users would rate the general mood of other content, and could then filter accordingly. If you’re not interested in the angry comments on an article, but would like to read some level-headed conversation, this would sort through the mess for you.

If you’re working on these problems already, I’d love to hear about it, so leave a comment or get in touch!