Just Say No to Trivia
People love trivia, and they think that knowledge of trivia ahould be rewarded. The success of quiz shows is enough to tell us that. Granted, many of the questions on quiz shows do deal with topics which it is important to know about, but many, if not most, are about topics which hardly anyone needs to know about. They are generally considered to be about knowledge of trivia.
We worship facts. If something is true, we believe that we are better off to know it. We even arrange for people to win large sums of money if they can demonstrate more knowledge of facts than other people.
Our fascination with facts leads us to attribute marvellous qualities to them. Chiefly we believe that all facts are useful and informative, and that the more facts we know the better off we are. Part of the lure of data mining is that it seems as if it will let us cope with more facts than we ever could before.
But not all facts are useful and informative. Surely, you may protest, it's better to know things than not to know them. However, if facts are not relevant to your daily life, then you're better off not to waste your time learning them. For example, you would not try to find your way in an unfamiliar city by first learning the name of every street in it.
Many people, though, have the idea that they can acquire knowledge about a topic by learning every fact connected with the topic that they can. Inevitably many of the facts they collect will be of little or no utility.
A more effective approach is to acquire information rather than facts. Information consists of data which are useful in making a decision. For example, if you're thinking of moving from another country to Canada, knowledge of job opportunities, form of government, climate, language, and so on are directly relevant to the life you will lead should you decide to settle there. Knowing the name of the prime minister of Canada in 1893 is not important.
The same approach is effective when doing things like deciding what to put in databases. If you want to include census data in your database, for example, the most effective procedure is to collect only the census data that are relevant to the decisions you want to make by referring to the database. That means you have to find out whether any of the census data available to you tell you anything useful and new about the population, but the time you spend doing that will cost you less than the consequences of making a decision based on irrelevant facts.