Data vs Information

A lot of organizations and individuals use the term 'data' synonymously with information or intelligence. I would like to add some clarification.

(Note: Data is the plural of Datum - Thank you Dr Mattingly, and the VVS school system)

A datum is a fact (so data would be several facts, related or unrelated.)

For our first 'mental exercise', if I roll 1 die, what is the probability of rolling a 6?

How many automatically thought (possibly derisively) 1 in 6?

Did you 'assume' I must have meant a 6-sided die? Any role-playing gamer knows there are 4-sided dice (so the probability is 0) or even 20-sided dice. A hidden bias was the assumption that that my question meant a 6-sided die. But I never stated that. Some assumed this.

(For this next mental exercise, I will be using Amazon - Because I purchase many books from them and they do their own 'analytics', the Author Turtledove - because I enjoy his writing, and these should not be interpreted as criticism of other vendors)

A lot of literature and organizational effort goes into the field of 'Data analysis' or 'data engineering'. When I was in the military, we were taught Data is an element of information. Intelligence is actionable after you correlate multiple data points.

So, given the fact I recently purchased a book by Harry Turtledove (Science Fiction, Alternate History) on Amazon (and enjoyed it immensely). how did Amazon use that data [point]?

Amazon, based on my purchase, recommended other books by Harry Turtledove and Science Fiction (content-based recommendation). They recommended books by Robert Conroy based on the fact others who read Turtledove have read Conroy (collaborative filtering). If Amazon performed analysis of the content they would have noticed both authors have a strong military theme (e.g. what of a different side had won a war or battle) (similarity Matching).

All of these methods of analysis require more than one data point. They require more than one book, more than one author, and/or more than one reviewer.

Given the fact I also purchase books (from Amazon, Barnes and Noble, and Books a Million, both online or in brick and mortar stores) on IT related topics. Amazon's analysis is actually Information Analysis and has, as I am about to point out, several gaps or flaws.

I also purchase books at Fantasy and Science Fiction Conventions I attend. I typically see authors selling their books in booths at the conventions. I have purchased a number of Star Trek fan fiction books. Amazon, being unaware of these purchases fails to recommend them to me. In fact, Amazon is also unaware of any purchases I make in any brick-and-mortar stores (bookstores or second-hand- bookstores). These genres also fail to make it to my Amazon Recommendations.

My Girlfriend, knowing my enjoyment of books by Turtledove has purchased as gifts, some books by Turtledove. Amazon, being unaware of this keeps recommending books I have and have already enjoyed. This behavior also occurs if I have purchased the books elsewhere. Books I have purchased for others does not mean I am now fond of that author or genre myself.

I am not saying "Avoid the use of Analytics" (especially and I have myself used the term 'Data Engineer' in describing my role in organizations). My advice is twofold:

Recognize the fact that you are actually analyzing 'information' based on multiple data points. Given one data point, any predictions will be highly inaccurate,

Recognize your analysis has hidden flaws and biases. Actively look for those biases. Blindly assuming your predictions must be accurate will lead to inaccurate decisions and potentially flawed actions.