In this survey, each worker responded to the following questions:
- What made you happy in the last 24 hours?
- Which events made you happy in the last three months?
We applied machine-learning algorithms to extract key insights from these happy moments. The conclusion of this study tells a lot about who we are, what makes us happy and the power of classification algorithms to profile who we are.
What is a ML classification algorithm? Essentially, classification means that the algorithm determines the "right categories of a new observation on the basis of a training dataset containing observations whose category membership is known". Let's look at the following example:
- Write a small text describing what makes you happy.
- Using the known training dataset, the ML classification algorithm will determine with a 65% to 80% accuracy from which country you are, if you are a man or woman, if you have kids. And even if you are married, single or divorced.
With a precision of 65% to 80%, ML classification algorithms learn from historical training data, and predict who we are, simply by analyzing a couple of sentences we write.
Imagine how such a powerful tool can be used - and misused - by marketing firms, political parties or employers. On a more constructive note, how can you think of an application in our workplace to transform the way we run our business?
- Right graph below: the top five keywords quoted by women as their source of happiness, are husband, daughter, son, home and family (sorted by frequency). The next one is friend; job first comes in the 8th place. Night also makes it to the top 10 list of most frequent keywords.
- Left graph: conversely, the most frequent words used by men in their happiness moments start with work and friends, then wife, home, dinner and girl friends... Family is (only) the 9th most frequent keyword, slightly behind morning in the 7th place.
Some thought-provoking outcome worth a discussion at the Thanksgiving dinner table tonight, isn't it? What do you think about it? Please share your thoughts and comments.
This analysis is based on the so-called " term frequency" in the entire corpus of words, after removing punctuation and general words.
What do we really mean by personal"achievements"?
- Work, job and money (in this order) are the top three keywords we use to describe our achievements.
- Being "well" only shows up in the tenth place, and only a few respondents make mention of helping other, or making a difference in their community. Yes, this is the society we currently live in!
In the category "nature", the top keywords are weather, outside, walk, rain andmorning.
This is only the tip of iceberg, there are many more categories and interesting insights. The details of this study are published in a IPython Notebook in the Kaggle data science platform under the Apache 2.0 open source license.Please scroll down to the bottom and look at the other graphs, I am sure that you will have fun deciphering the "human nature" of our society.