Smarter Data Science

Оглавление

Cole Stryker. Smarter Data Science

Table of Contents

List of Illustrations

Guide

Pages

Praise For This Book

Smarter Data Science. Succeeding with Enterprise-Grade Data and AI Projects

About the Authors

Acknowledgments

Foreword for Smarter Data Science

Epigraph

Preamble

Why You Need This Book

What You'll Learn

CHAPTER 1 Climbing the AI Ladder

Readying Data for AI

Technology Focus Areas

Taking the Ladder Rung by Rung

THE BIG PICTURE

Constantly Adapt to Retain Organizational Relevance

ECONOMICALLY VIABLE

Data-Based Reasoning Is Part and Parcel in the Modern Business

LEARNING

Toward the AI-Centric Organization

SCALE

Summary

CHAPTER 2 Framing Part I: Considerations for Organizations Using AI

Data-Driven Decision-Making

Using Interrogatives to Gain Insight

NOTE

The Trust Matrix

The Importance of Metrics and Human Insight

THE ZACHMAN FRAMEWORK

Democratizing Data and Data Science

DEMOCRATIZATION

Aye, a Prerequisite: Organizing Data Must Be a Forethought. NOTE

NOTE

NOTE

NOTE

Preventing Design Pitfalls

ARCHITECTURE AND DESIGN

Facilitating the Winds of Change: How Organized Data Facilitates Reaction Time

NOTE

MUTABLE

Quae Quaestio (Question Everything)

QUESTIONING

Summary

CHAPTER 3 Framing Part II: Considerations for Working with Data and AI

Personalizing the Data Experience for Every User

NOTE

NOTE

WATER

Context Counts: Choosing the Right Way to Display Data

CONTEXT

Ethnography: Improving Understanding Through Specialized Data

DRILLING DOWN

Data Governance and Data Quality

The Value of Decomposing Data

Providing Structure Through Data Governance

Curating Data for Training

Additional Considerations for Creating Value

STANDARDIZATION

Ontologies: A Means for Encapsulating Knowledge

SEMANTICALLY DISAMBIGUATE

Fairness, Trust, and Transparency in AI Outcomes

NOTE

NOTE

ETHICS

Accessible, Accurate, Curated, and Organized

CURATED

Summary

CHAPTER 4 A Look Back on Analytics: More Than One Hammer

Been Here Before: Reviewing the Enterprise Data Warehouse

NOTE

NOTE

NOTE

NOTE

NOTE

NOTE

NOTE

A RELATIONSHIP IS NOT JUST A LINE BETWEEN OBJECTS

Drawbacks of the Traditional Data Warehouse

NOTE

NOTE

NOTE

NOTE

NOTE

THE LOGIC BEHIND A BEST PRACTICE

Paradigm Shift

NOTE

ANY VOLUME IN ZERO SECONDS

Modern Analytical Environments: The Data Lake

By Contrast

NOTE

NOTE

NOTE

Indigenous Data

Attributes of Difference

RAW DATA

Elements of the Data Lake

BIG DATA QUALITY

The New Normal: Big Data Is Now Normal Data

Liberation from the Rigidity of a Single Data Model

Streaming Data

Suitable Tools for the Task

Easier Accessibility

Reducing Costs

Scalability

Data Management and Data Governance for AI

FACTORS

Schema-on-Read vs. Schema-on-Write

AN UNDERLYING METAMODEL

Summary

NOTE

NOTE

CHAPTER 5 A Look Forward on Analytics: Not Everything Can Be a Nail

A Need for Organization

NOTE

NOTE

NOTE

The Staging Zone

The Raw Zone

The Discovery and Exploration Zone

The Aligned Zone

The Harmonized Zone

The Curated Zone

DATA RICH, INFORMATION POOR

Data Topologies

NOTE

Zone Map

Data Pipelines

NOTE

Data Topography

MISGUIDED TENETS

Expanding, Adding, Moving, and Removing Zones

LEAF ZONES

Enabling the Zones

Ingestion

Data Governance

Data Storage and Retention

NOTE

Data Processing

Data Access

Management and Monitoring

Metadata

WHITE BOX, GRAY BOX, BLACK BOX

Summary

CHAPTER 6 Addressing Operational Disciplines on the AI Ladder

A Passage of Time

NOTE

NOTE

ADAPTIVE OVER AGILE

Create

Stability

NOTE

Barriers

Complexity

REDUCTION

Execute

Ingestion

NOTE

Visibility

Compliance

MVP

Operate

NOTE

Quality

NOTE

Reliance

Reusability

ADAPTIVE

The xOps Trifecta: DevOps/MLOps, DataOps, and AIOps

DevOps/MLOps

DataOps

AIOps

ADAPTIVE

Summary

CHAPTER 7 Maximizing the Use of Your Data: Being Value Driven

Toward a Value Chain

NOTE

NOTE

Chaining Through Correlation

NOTE

Enabling Action

Expanding the Means to Act

IT'S ALL JUST METADATA

Curation

NOTE

FIT FOR PURPOSE

Data Governance

WAIVERS

Integrated Data Management

Onboarding

Organizing

NOTE

Cataloging

Metadata

Preparing

Provisioning

Multi-Tenancy

FEATURE ENGINEERING

NOTE

Summary

CHAPTER 8 Valuing Data with Statistical Analysis and Enabling Meaningful Access

Deriving Value: Managing Data as an Asset

NOTE

NOTE

ATTENTION TO DETAIL

NOTE

NOTE

NOTE

An Inexact Science

NOTE

FROM THE BEGINNING

Accessibility to Data: Not All Users Are Equal

HIDDEN BY NECESSITY

Providing Self-Service to Data

AVOIDING VAGUE OR AMBIGUOUS METADATA

Access: The Importance of Adding Controls

LYING

Ranking Datasets Using a Bottom-Up Approach for Data Governance

DATA QUALITY APPLIES

How Various Industries Use Data and AI

BOUNDARIES

Benefiting from Statistics

NOTE

MISAPPLIED

Summary

CHAPTER 9 Constructing for the Long-Term

The Need to Change Habits: Avoiding Hard-Coding

NOTE

NOTE

NOTE

Overloading

NOTE

Locked In

SIMPLE ISSUES MAY NOT ALWAYS BE SIMPLE TO MITIGATE

Ownership and Decomposition

Design to Avoid Change

OSAPI

Extending the Value of Data Through AI

NOTE

TIME IS AN INTANGIBLE ASSET

Polyglot Persistence

NOTE

MODELS DEPLOYED AS MICROSERVICES

Benefiting from Data Literacy

Understanding a Topic

Skillsets

NOTE

It's All Metadata

NOTE

NOTE

The Right Data, in the Right Context, with the Right Interface

NOTE

ASCERTAINING CONTEXT

Summary

CHAPTER 10 A Journey's End: An IA for AI

Development Efforts for AI

RETRAINING

Essential Elements: Cloud-Based Computing, Data, and Analytics

RESILIENCY

Intersections: Compute Capacity and Storage Capacity

SUSTAIN

Analytic Intensity

xPU ACCELERATION

Interoperability Across the Elements

NOTE

NOTE

A USE CASE

Data Pipeline Flight Paths: Preflight, Inflight, Postflight

A USE CASE (CONTINUED)

Data Management for the Data Puddle, Data Pond, and Data Lake

NOTE

A POTENTIAL BABEL FISH

Driving Action: Context, Content, and Decision-Makers

EXPLAINABLE AI

Keep It Simple

ALTERNATIVES TO COMPLEX DATA SECURITY PROFILES

The Silo Is Dead; Long Live the Silo

NOTE

THE BODY AS A MYRIAD OF SILOS

Taxonomy: Organizing Data Zones

THE BODY AS A MYRIAD OF SILOS: SPECIALIZATION

Capabilities for an Open Platform

Summary

Appendix Glossary of Terms

Index

WILEY END USER LICENSE AGREEMENT

Отрывок из книги

The authors have obviously explored the paths toward an efficient information architecture. There is value in learning from their experience. If you have responsibility for or influence over how your organization uses artificial intelligence you will find Smarter Data Science an invaluable read. It is noteworthy that the book is written with a sense of scope that lends to its credibility. So much written about AI technologies today seems to assume a technical vacuum. We are not all working in startups! We have legacy technology that needs to be considered. The authors have created an excellent resource that acknowledges that enterprise context is a nuanced and important problem. The ideas are presented in a logical and clear format that is suitable to the technologist as well as the businessperson.

Christopher Smith, Chief Knowledge Management and Innovation Officer, Sullivan & Cromwell, LLC

.....

Advanced analytics, including AI, can provide a basis for establishing reasoning by using inductive and deductive techniques. Being able to interpret user interactions as a series of signals can allow a system to offer content that is appropriate for the user's context in real time.

To maximize the usefulness of the content, the data should be of an appropriate level of quality, appropriately structured or tagged, and, as appropriate, correlated with information from disparate systems and processes. Ascertaining a user's context is also an analytical task and involves the system trying to understand the relationship between the user and the user's specific work task.

.....