Lies, Damned Lies and LinkedIn Trends

There are observations, there are trends and there are inferences which are there just because we can- “which are full of sound and fury, but signify nothing” ;o)

The reason for this blog is a set of inferences from LinkedIn Blog on the job promotions

Generally good work and the team under DJ are the best …

Yet, sometimes people go overboard …

One inference they are making is that there are fewer January promotions in younger generations …

  • While the covariance of promotions and January is interesting, I cannot conclude that they are correlated. It needs more context – for example: one factor, I can think of , is that compare the Fiscal Year of the companies (especially the ones) who have CY = FY and then see if the raises correlate with the companies who have CY (Calendar year) = FY (Fiscal year). May be there are more companies with CY=FY now than before !
  • I am sure one can make a chart showing people under 30 should not ask for promotions in months with more than 30 days, and while that might be a valid inference … I will let you conclude whether this is domain causality or accidental causality ;o)
  • This where similarity coefficients/symmetric-assymmetric objective measures like Jaccard/Tanimoto Coefficient, Gini Index and the Cosine Similarity, that test for presence and absence equalities, might help (I am still an amateur at this). Only in the context of these metrics can we make valid inferences …
  • Before concluding that 30 year olds have a less chance of getting promoted in January, may be it is because the companies who have more 30 year olds have the FY ending on July
  • Also we need a query which shows the promotions months for different age groups and that might tell us which months are best for which age range. This again need to correlate with the CY/FY of the companies and other trends
  • Also remember that this is a self-selecting population – the LinkedIn sample might not represent the general employment market

Anyway, I know I shouldn’t be very critical about this. I have been following the LinkedIn blogs and the folks do a good job in generating inferences …

But I also want to make sure they don’t go overboard and start creating trends with irrational exuberance …

While correlating promotions and months is hardly an inference of gigantic proportions, I do want to remind the folks that

With Big Data comes big responsibilities

Just because one can run a query doesn’t mean …

I rest my case, what says thee ?




2 thoughts on “Lies, Damned Lies and LinkedIn Trends

  1. Pingback: Tweets that mention Lies, Damned Lies and LinkedIn Trends « My missives --

  2. Pingback: Strata 2011 Keynote logs – Stardate 64587.1 « My missives

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s