Data Analysis as a Form of Story Telling

The year is 1854.  London is in the middle of yet another cholera outbreak.  There had been one in 1832, and another five years earlier in 1849 that killed 14,137 Londoners.  Mostly women and children.[1]  Cholera is a nasty thing.  People suffering from cholera experience watery diarrhea that can lead to severe dehydration.  This results in sunken in eyes, cold skin, and turning blue.  In addition, it may also cause vomiting and muscle cramps and as seen above, death.  The germ theory had not been established yet – Louis Pasteur would not propose it until seven years later in 1861 – and instead people held to the miasma theory, which held that somehow “bad air” was to blame.  It certainly was not the slaughter houses and grease boiling dens lining the streets, or the overrunning cesspools underneath the cellars.  Doing what governments do, the London authorities wisely dealt with the cesspool problem by pumping the sewage into the Thames.  Problem solved.  Miasma theory is, incidentally, the same theory that proposed one could become obese by smelling too much food.  Thankfully, none of this sounded quite right to an epidemiologist named John Snow.   Dr. Snow decided to do something quite radical for the time.  Rather than fearing sinister “bad air”, Dr. Snow wanted to approach the problem with data.  He decided that he needed to collect data, and then analyze that data. When Dr. Snow had gathered the information, he did something else that was revolutionary.  He put his data onto a map, creating a visual chart of his data.  Dr. Snow was an early data analyst.  Was Dr. Snow, the budding data scientist successful in his efforts?  We’ll come back to Dr. Snow and the dear old London at the end of this paper.