About Time Series

What can I say about time series data that hasn't already been said?   Quite a lot, it turns out.  At Epinomy, our attention has been quite well focused on a particular kind of structured data known as "time series".   Who knew these things were so ubiquitous?

We started out with the idea that you can apply search techniques like those  used by Amazon.com and Google to find any old structured data.   By structured data, I mean tables and relational databases as opposed to unstructured data like documents, e-mails, tweets and other textual data sources.   

As part of our development efforts, it became clear that there is a particular kind of data set that is in use by just about every large organization on the planet.  Time series are everywhere.   The operational data of any organization has a time dimension associated with it, and that's what a time series is:  a list of measurements of some quantity that changes over time.

Just off the top of my head, here are a half dozen verticals and some of the time series that they track:

Financial Services

  • Stock Prices 
  • Interest Rates
  • Exchange Rates

Pharmaceuticals

  • Pharmacokinetics
  • Adverse Events
  • Dosing
  • Sales
  • Production

etail

  • Clickstreams
  • Sales
  • Purchases
  • Margin

 

Manufacturing

  • Units Produced
  • Defects
  • Units Shipped
  • Net Margin

Health Care

  • Patient Vital Signs
  • Patient Billing History
  • Interventions

Government

  • Contractor Hours
  • Epidemiology
  • Logistics
  • Demographics
  • Law Enforcement

That's literally 3 minutes of thinking - there are hundreds more applications for time series search.

What's your favorite time series?

 

Google+