In a post written the 27th of October on non-confirmatory exploratory data analysis I expressed concern about the reliability of its results for supporting decisions and actions. I do think that if properly/competently done its results can be reasonably reliable. But what does properly done mean.
If improperly done non-confirmatory exploratory data analysis reminds me of Munchhausen’s Journey to the Moon. I am also reminded of Buddha’s parable about blind Indians trying to describe an elephant.
“The Buddha related his famous parable. ‘Once here in Savatthi, the king called a certain man and said: “Assemble together in one place all the men in Sāvatthi who were born blind.” Having done as the king commanded, the king then said to the man, “Now show the blind men an elephant.” Again the man did as the king commanded, saying to each as he did, “Oh blind man, this is an elephant and this is its head. This is its ear. This is its tusk. This is its trunk. This is its body. This is its leg. This is its back. This is its tail. This is the end of its tail.” This having been done the king addresses the blind men saying, “Have you seen an elephant?” and they replied, “We have sire.” “And what is an elephant like?” he asked. And the one who had touched the head said, “An elephant is like a pot.” while the one who had touched the ear said, “An elephant is like a winnowing basket.” The one who had touched the tusk said, “An elephant is like a plough pole” while the one who had touched the trunk said, “It is like a plough.” The one who had touched the body said, “It is like a granary,” and the one who had touched the leg said, “It is like a pillar.” The one who had touched the back said, “It is like a mortar”, the one who had touched the tail said, “It is like a pestle” while the one who had touched the end of the tail said, “An elephant is like a broom.” Then they began to quarrel saying, “Yes it is!” “No it isn’t!” “An elephant is like this!” “An elephant is like that!” until eventually they began fighting with each other.’ Having told this story, the Buddha summed up its meaning in a terse little verse – ‘Some monks and priests are attached to their views and having seized hold of them they wrangle, like those who see only one side of a thing.’ -”
The blind Indians might be analogous to data analysts who are engaged in improperly done basic exploratory analysis. Seeing Indians might then be analogous to data analysts who are engaged in confirmatory exploratory data analysis. Presumably they can describe how the elephant really looks. But how close does their description really come to the real look of the elephant?
This question leads to the epistemological question: “How does reality really look”?