PREDICTIVE ANALYSIS Design of Experiments TYPES OF DATA

PREDICTIVE ANALYSIS Design of Experiments

TYPES OF DATA ANALYTICS • Descriptive Analytics • Diagnostic Analytics • Predictive Analytics • Prescriptive Analytics

DESCRIPTIVE ANALYTICS • Quite literally, a description of the data • Example – A company sold 90, 000 widgets in the first quarter of the year and another 124, 000 widgets in the second quarter of the year • Descriptive Analytics tell us how many widgets were sold in the first and second quarter of the year, and by virtue, the first half of the year

DIAGNOSTIC ANALYTICS • This is the “why” behind the numbers, but it is still a “what happened” style of analysis • Using our example – A data analyst can determine if certain sales and marketing efforts contributed to the number of sales. Was there a targeting marketing campaign? Was there a regional increase? Were new salespeople hired? Did the company change the incentives for the sales department?

PREDICTIVE ANALYTICS • Here we look at the data and predict what might happen next • Prediction with 100% certainty is not the goal (think weather) • Thoughtful analysis of existing data points help ‘predict’ a likely occurrence

PRESCRIPTIVE ANALYTICS • The next step in the sequence is to make a recommendation for next steps based on information determined in the predictive analysis phase • Version 2. 0 etc.

LET’S GIVE IT A SHOT!

HOW MANY VARIATIONS? • There are 3 configurations of 4 different variables • Variables are noted as A, B, C & D • Configurations are noted as A 1, A 2 & A 3 • 3 to the 4 th power = 81 possible configurations • Is it a good idea to test all possible configurations? • Let’s propose a logical subset to examine

DATA SUBSET

VOLUNTEERS! • Nine sample students • One official measurement person • One organizer/data collector

DESCRIPTIVE ANALYTICS • REMEMBER, it’s a description of the data • How far did the paper airplanes go? • Distances • Average Distance • Etc.

FLIGHT DATA TEMPLATE

DIAGNOSTIC ANALYTICS • REMEMBER, this answers the ‘why’ question • What happened? • Launch angle • Strength of thrower • Quality of construction • Measure from what point (front vs. back) • Etc.

DESIGN OF EXPERIMENTS • Controllable input factors – ‘X’ variables • Design variables A, B, C & D • Uncontrollable input factors – ‘Y’ variables • Predicted variables and those that were uncovered during the initial experiment • Responses – Measured output

PREDICTIVE ANALYTICS • REMEMBER, here we look at the data and predict what might happen next • Look for outliers • Look for correlation

INITIAL FLIGHT DATA

VARIABLE ‘A’ & ‘B’ ANALYSIS

VARIABLE ‘C’ & ‘D’ ANALYSIS

PRESCRIPTIVE ANALYTICS • REMEMBER, the next step in the sequence is to make a recommendation for next steps based on information determined in the predictive analysis phase • What should we test next?

NEW EXPERIMENTS

RESULTS