How to stop reporting and start analyzing
I’ve worked for years in Business Intelligence (BI). From an analysts to a director, my hands have been in every part of this profession. But I have to make a confession, most of the time I wasn’t doing classic BI.
In fact, almost no BI as it currently exists is similar to the original meaning. In the beginning, way back in 1865 when it was coined by Richard Millar Devens, Business Intelligence referred to using information from the business environment to gain a competitive advantage. Today this would be something like applying macro-economic data or competitors’ financial reports to make sales projections.
Unfortunately, modern BI consists almost entirely of using internal business data. How many people reading this use any type of external data in their day to day business activities? Not many I would guess, although the ones that do are certain to send me tweets telling me I’m wrong… Regardless, the idea of incorporating data from outside the business into reports or analyses has moved away from modern BI and into the hands of Analytics or Econometrics Departments.
Why? Probably because people think that data is too hard to get, or the internal data is too easy to get and no one seems to be asking for anything else, so why bother?
And that’s the problem, people are not asking BI for external data because they don’t know BI can do it, they think it’s too hard to do! But here’s the thing, it’s not, if you know where to start.
Here are a few tips on how to bring back classic BI with modern technology!
1. Learn how to use an API
“Application Program Interface”, API, sounds scary doesn’t it? Most of the time, they are not. Typically they are as easy and knowing what to type in a URL so you get data back. Once you know the URL, you can use that in TIBCO Spotfire, Tableau, even Excel to automatically get your data. See my article on using the Census API in TERR for some details.
If you want to get really advanced, Codecademy has a bunch of tutorials on how to use APIs: https://www.codecademy.com/learn
Once you understand how APIs work, go to Quandl.com and start pulling in all the data you’ll ever need.
2. Learn some Statistics
Come on, we all took business statistics in college, this stuff isn’t hard! All you need to learn is how to do correlation and regression and you’ll be well on your way to comparing all the data you’re getting from APIs against your business data!
Go through Khan Academy’s Stats course for an excellent primer: https://www.khanacademy.org/math/probability
3. Learn R to bring it all together
Okay, learning a new programming language is actually pretty hard, particularly R. It’s a weird language designed 50 years ago to be executed on a command line. However, if you spend a few weeks with it, even if you’ve never programmed before, I guarantee you will find yourself doing things you never thought possible with your existing BI tools.
Why is that? Because more and more BI tools are adding R functionality. R is a completely free math and statistic tool, unlike say SAS. It doesn’t add to licensing costs and it has extended functionality that covers effectively every mathematical technique every invented.
You want to analyze the sentence structure of Tweets? There’s a package for that.
You want to do time-series forecasting? There are a whole bunch of packages for that!
But if you want to get started, grab yourself a free open source edition of Rstudio.
And go through this tutorial.
With the ability to get data, and do statistical analysis on it with R, you can start applying your new skills. TIBCO Spotfire and Tableau both have robust R integration built in to their standard platform. If you would like to know more, check out their information online or contact a Syntelli data wizard today!
