Predictive Analytics for Practitioners

Explore a comprehensive overview of Predictive Analytics in this 56-minute conference talk from the PASS Data Community Summit. Gain insights into four commonly misunderstood topics: data preparation, sampling, algorithm strengths and weaknesses, and model accuracy assessment. Dive into the CRISP-DM Process Model, covering business understanding, data understanding, data preparation, modeling, and deployment steps. Learn about classifier decision boundaries, ROC curves, and the differences between Predictive Analytics and Data Science. Discover what it takes to become a Predictive Modeler and follow the Analyst's Journey through this informative session led by Dean Abbott.

Syllabus

Intro
Technical Assistance
What do Predictive Modelers do? The CRISP-DM Process Model
CRISP-DM: Business Understanding Steps
Objective's
Data Understanding Steps
Source Data
Data Preparation (Conditioning) Steps
Data Size
Modeling Steps
Sampling
Classifiers Find Different Decision Boundaries
Assess Models: ROC Curves
CRISP-DM Step 6: Deployment Steps
Model Results after Deployment
What is Predictive Analytics? Simple Definitions
Predictive Analytics vs. Data Science
What Degree Does it Take to Be a Predictive Modeler?
The Analyst's Journey