TIBCO Spotfire:A Comprehensive Primer(Second Edition)
上QQ阅读APP看书,第一时间看更新

Insights from details visualizations

Now that we have created a useful details visualization, what conclusions, insights, and findings can be drawn from it? I'll share some of mine with you—see if you can find more of your own:

  1. Here, I have selected all the data. The age of the passengers is mostly normally distributed, but with a peak at the lower age range (there seemed to be a lot of babies on board):

Normal distribution
A lot of data is normally distributed, particularly measured data—that is, data that has been recorded as a result of measured observation of real-world events or phenomena. Measuring the age of a population will nearly always result in some form of normal distribution. Blood pressure data in a patient population is usually normally distributed. I'm sure you can think of many other examples.
The normal distribution has informally been called the bell curve. This is because the curve looks like it's bell-shaped, with an enlarged middle, tailing off at each end.
  1. Very young babies stood a good chance of survival, as did most young children, with the exception of about 3-year-old children.
  2. Children from about 9 years old to 14 years old didn't fare too well, sadly.
  3. There were very few passengers aged 60-80, but most of them did not survive. There is a lone exception—an 80-year-old male that did survive—good for him!
  4. Compare the visualization for males versus females by selecting all the males, and then selecting the females.
    This visualization shows the female passengers:

This histogram shows the survival rates of the male passengers only:

  1. How much more depressing is the male survival histogram than the female? We already know that a much higher proportion of men didn't survive the disaster, but this visualization is very telling. With the exception of a few peaks at various age ranges, your chances of survival as an adult male were very slim.

Feel free to experiment by selecting other parts of the data in the original bar chart and seeing how it affects the histogram—see if you can gain any additional insights—I have just named a few!