Maandelijks archief: oktober 2011

Good bye

Today is the sad day that Steve Jobs died. I want to share a personal story about how his words, simple as they are, have motivated me recently.

I’ve watched his Stanford speech, which I will embed below, several times in the past year. But like most people who watch this, I merely loved it, and didn’t act on it. Until this summer. When I was on holiday, I made a note in my iPhone, which is paraphrasing Steve, I think from the D8 interview.

Going to work should be motivated by a will to do something amazing, to build great products.

I realized that I wasn’t doing anyone a service in my role at my employer. So after I got back from vacation and gathered enough courage, I handed in my resignation. Coincidentally, it was on the same day that Steve sent his famous letter to the board, informing them that he could no longer act as CEO of Apple.

And tonight my farewell party is planned. Coincidentally, on the very sad day that Steve has passed away. I hate saying good bye, so today is a good exercise in that.

Steve’s words motivated me to take my decision. In that sense, he will live on, because my life is affected by him. And this is just one little story. I hope he will motivate you too. 

Your time is limited, so don’t waste it living someone else’s life.

You should really watch his Stanford speech. Here it is. 

First talk on CSN11 – Sentiment analysis

According to the Condorcet theorem, bigger crowds do better. If each individual independently decides with the same probability, the collective reaches ‘the right’ decision fast. If the individuals can connect, the collective behavior can become very rich and complex. See Linux, wikipedia.

But on social networks, there is no collective ‘goal’ besides connecting, recreation. Still, the chatter can predict the box office receipts. And Google search terms shows where people have flu, better than any medical database. It is also shown that happiness and loneliness is contagious; if many people in your network are lonely, you feel more lonely too.

His research focuses on sentiment, by extracting indicators from text. And predicting the future from that. This can work well, if applied on large numbers, see http://wefeelfine.com and Facebook gross happiness index.

McNair defines 72 terms, that map to 6 independent dimensions. These terms are applied to tweet texts, resulting in 6 charts. It turns out that about 8% of all tweets show emotional content. Charts of Election08 shows dips and peaks that make sense; energetic on election day, for example.

The Dow Jones industrial average appears to correlate with the Calm measurement, but with a lag of about 4-5 days. It can be used to predict its future value, with an accuracy of 86%, whatever that may mean.. See http://arxiv.org/abs/1010.3003 for his paper.

He also studied mood contagion. Look at twitter users who follow each other, and then the use of emoticons. People follow others with similar traits, assortative network. He shows that sad people connect to sad people, and happy to happy people. His recommendation is to connect to happy people, and to disconnect from unhappy ones. šŸ™‚

He concludes with the question if we can unleash a happiness virus on the network, would that work? Or push another mood?

Nice talk by jbollen@indiana.edu. The slides of his presentation can be found at http://www.slideshare.net/KREMCSN/csn11-keynote-johan-bollen