Data Analysis

Large amounts of data often have buried within them secret relationships, causality, and predictable events. We conduct data analysis in order to tease out this information. We search for the connections between variables that are thin, but where multiple relationships accrue toward something that is tangible and useful. In business, there may be relationships between advertising and sales, R&D and long-term profits, executive skills and market share. Academics search this data for one more undiscovered relationship. They need something new to be able to create a publishable journal paper.

But what happens when the data does not contain any relationships? What happens when all of the relationships and secrets have been uncovered? Then what? Does that branch of academia evaporate or redefine itself?

I am studying the Top 100 companies that invest in R&D in the US. I have plotted variable in 18 different pairs. Most of these graphs show no relationship within the data. If anything, it says, “large companies spend more that small companies” or “technology companies spend more than industrial companies”. This is not exactly revolutionary information. I think we could have guessed that. The next step is to divide the companies into different groups – like High Sales, High Capital, High R&D, Computer Technology, Biotech, Pharma, etc. This may allow us to see industry trends rather than country trends. Perhaps there is less commonality regarding R&D spending by country than there is R&D spending by industry.

This has been a good sounding board. I think that is something to pursue.