How To: Solve Business Problems with Data

Two approaches I have found helpful to understand any business situation and choose the correct techniques to analyze data.

1.      Cross Industry Standard Process for Data Mining (CRISP-DM)

2.      Predictive Methodology Map

CRISP-DM

This framework was originally developed by data miners to generalize the common approaches to defining and analyzing a problem.

The framework is made up of 6 steps:

1.      Business Issue Understanding

2.      Data Understanding

3.      Data Preparation

4.      Analysis/Modeling

5.      Validation

6.      Presentation/Visualization

 Methodology Map

The methodology map is a guide to determine the appropriate analytical technique(s) to solve a particular business question.

The map outlines two main scenarios:

1.      Data analysis

2.      Predictive analysis

Data analysis refers to standard approaches of blending together data and reporting on trends and statistics and helps answer business questions that involve understanding more about the dataset such as "On average, how many people order coffee and a doughnut per transaction in my store in any given week?"

Predictive analysis helps businesses predict future behaviour based on existing data such as "Given the average coffee order, how much coffee can I expect to sell next week if I were to add a new brand of coffee?"