Note: Twitter Data Shows When We’re Happy, Sad, Hungover


Last week Mashable reported that Twitter had gathered data about key words used in tweets which indicated the moods and behaviour of people on each day of the week for each month of the year.

Essentially they have aggregated the frequency of the following terms:

  • “feel happy”
  • “feel sad”
  • “hungover”
  • “late for work”

So, what did this reveal?

The most interesting finding to me is that December is a month of highs and lows. While Tuesdays in December often attract “feel happy” tweets, this is the month of the year when users are most likely to use the term “feel sad” in their update.

Why am I interested in this?

Part of the ‘Cheer up love’ care labels project which I have not really explored here is exchanging data from mobile apps with a wearable item like the care labels I have talked about.

So it could be possible that setting the wearable would add a ‘signature’ to your texts, emails or social media updates. This might be the combination that you have set the wearable to that day.

But taking it further, could the wearable item receive data from your texts, emails, social media updates or other quantified self apps or devices which would prompt the wearer to adjust their labels?

I’ve spoken before about using a mood tracking program such as MoodScope to recommend care labels for that day. MoodScope currently only records one mood score per day – but other apps could continue to inform your choice of wearable care labels on a much more regular basis.

First steps to generating a list of care labels



To get started I wrote a massive list of the kind of needs I thought people would feel the need to express using care labels. They fall into these general categories:


  • have a cup of tea and a chat
  • share a meal and talk


  • call me
  • text me
  • Facebook / Twitter / G+ me


  • handle with care
  • no questions
  • no difficult questions
  • smile at me
  • be patient with me
  • let me be alone
  • don’t take offence


  • ask me how I am
  • be listened to
  • to listen to you
  • smile at me
  • introduce me to a friend
  • include me in your plans
  • tell me the truth

Physical presence

  • just be there
  • share a hug
  • need some space
  • approach with caution
  • leave me alone
  • spend some time together one-to-one
  • spend some time together in a group


  • celebrate with me
  • a good night out
  • have a laugh
  • see some comedy
  • watch a film
  • go to the theatre
  • dancing / clubbing

As you can see there is some room for fluidity in these categories, with several suggested care labels fitting into more than one list. I suspect there may be a Venn diagram in there somewhere and will be having a look at that next.

Mood scores

Additionally, I came up with the possibility of adding a score of 0-100. This is what the mood-tracking website MoodScope allows you to. This level of detail would not be possible with care label beads that only allowed for four variations, but could be stripped down to:

  •  < 25
  •  25 – 50
  •  50 – 75
  •  > 75

Is there anything obvious I have missed? Please add your thoughts to the comments below.

You can read more about the thinking behind Care Labels For Humans on the About page.