Initiating a big data project can be overwhelming. It’s not just one more thing on your plate — it’s more like three plates spinning at once. Take step one and see how you can use big data to solve a problem by locating hidden patterns in large volumes of data.
First, let’s figure out if big data will work for you — in three simple steps — without the mess of broken dishes.
Step 1 — Define the problem
Step 2 — Figure out how much data you need
Step 3 — Measure the problem against the data
Hint – If you don’t know your dilemma, what data you have, or where you want to end up then you need to focus on these three steps before diving into a data analysis project.
What’s your headache?
Do you have a headache? By a headache, I mean a pressing business need that needs to be addressed. Sam does.
Sam is a director of Admissions as a mid-size university, in charge of retaining students.
For many years the university has focused on the overall freshman year experience – and enjoyed a record number of freshman returning year two.
Recently, however, there has been a decline in the number of freshmen returning their sophomore year. Sam needs to know why, and how to reverse the decline. He’s hoping big data can help.
Is a big data project right for Sam?
Yes, he has a pressing business need — one that would apply to any business seeking to keep existing customers.
If his first-year retention was down by 13% last year, and if that trend continues for two more years, enrollment will be only 66% of its previous level — a critical business need!
Sam needs to know what’s causing this trend. What if the college’s rapid growth, or lack of new programming that caused students to look elsewhere after their freshman year? Which types of students left in droves, and which types remained?
Let’s see if data mining can identify the answers. Please read the next blog post to learn how.