Part 1: Data Science on the IBM Cloud

Data Science - Field of applying intensive scientific methods to clean and analyse data to reveal facts and employ insights.  

Data - Recorded displays of information practically stored digitally and represented as numbers, text, images and graphs, audio, and video that get interpreted by specific methods to achieve specific goals. Data gets inputted to IBM Cloud to become organized and undergo analysis.   Define client product and service goals that will increase market shares and boost profits by undergoing digital innovation, research and development, marketing and advertising, and refinement. 

 

Enterprise Scale Data Science

AI Lifecycle Management 

1. Define project subject and goals.

2. Pursue optimization based on business Key Performance Indicators (KPIs), measurements of efficiencies a team is able to achieve to meet business goals focused on revenue and profit margins.  

3. Upholds fairness, explainability of data applition to forecasted business objective, and maintains privacy. 

4. Provide iterations for repetitions able to generate sequences of outcomes that achieve both fixed and dynamic goals.

 

Infrustructure 

1. Ensemble of multiple data sources, lineages of data, and streaming data that is supported in a foundation comprised by multiple stakeholders that provides management effieincy and the abliity to accomodate all contributors.

 

Personal Data Science Process

Gather data > Cleanse data > Feature engineering > Select and develop models > Deploy optimized model

 

Enterprise Data Science Process

Data lineage and source selection > Data cleansed and prepared, as regular data flow requires governance, > Features and models supported by explainability > Deployment for model monitoring of fairness and deviation > New findings continuously applied towoards improving business outcomes in following cycles of the process.

 

Cycle continuously repeats and undergoes regular monitoring and improvement. 

 

Cloud Modernization and the AI Ladder

AI Ladder - Framework for understanding the work and processes necessary to implement AI based solutions for large enterprises.

 

AI Ladder

4. Infuse operational AI throughout the enterprise at optimized performance.

3. Analysis wihile applying data to build and scale AI development, with integrity.

2. Organize and present data into a business-ready analytical solution. 

​1. Harvest data in simple and accessible forms.

 

AI Ladder gets modernized to embrace the value of data applicated in an AI and hybrid, multi-cloud world.

 

Information Archietecture Ecosystems

For application of AI Ladder, enterprises must provide relevant infrustructure.

​

"No Artifical Intelligence without Information Archietecture, No AI without IA."

​

Information Archietecture Ecosystems - Evolution to interconnecting Data Focus Areas for successful scale and infusion of AI in an enterprise.

 

Data Focus Areas, cocurring deployment can drain time and resources

 

1. Data Persistence, including data storage, physical and digital tools used to collect data.

2. Data Fabric, the organizing, cleaning, and governance of data.

3. Data Science, coded and codeless solutions used to visualize and develop models.

4. Open Source, exists with focus areas 1-3 that is able to scale from an indivudal to a global population. 

 

IBM Cloud Pak for Data - Unified solution for deployment from IBM servers to customer destinations.