AI Enterprise Workflow Specialization

Certificates: Course 1, Course 2, Course 3, Course 4, Course 5

About this Specialization

IBM AI Enterprise Workflow is a comprehensive, end-to-end process that enables data scientists to build AI solutions, starting with business priorities and working through to taking AI into production.

*from the Specialization landing page.

Course 1. Case Study - Data ingestion

The goal of this case study is to put into practice the important concepts from module 1. We will go through the basic process that begins with refining the business opportunity and ensuring that it is articulated using a scientific thought process.

The business opportunity and case study was first mentioned in Unit 2 of module 1 and like the AAVIAL company itself these data were created for learning purposes. We will be using the AAVAIL example as a basis for this case study. You will be gathering data from several provided sources, staging it for quality assurance and saving it in a target destination that is most appropriate.

Case study overall objectives

1. Gather all relevant data from the sources of provided data
2. Implement several checks for quality assurance
3. Take the initial steps towards automation of the ingestion pipeline

Course 2. Case Study - Multiple Testing

When we perform a large number of statistical tests, some will have $p$-values less than the designated level of alpha (e.g. 0.05) purely by chance, even if all the null hypotheses are really true. This is an inherent risk of using inferrential statistics. Fortunately, there are several techniques to mitigate the risk.

We are going to look at the 2018 world cup data in this example.

The case study is comprised of the following sections:

1. Data Cleaning
2. Data Visualization
3. NHT
4. Adjust NHT results for multiple comparisons

Data science work that focuses on creating a predictive model is perhaps the hallmark of the field today, but there are still many use cases where inferential statistics are the best tool available. One issue with statistical inference is that there are situations where performing multiple tests is a logical way to accomplish a task, but it comes at the expense of an increased rate of false positives or Type I errors.

In this case study you will apply techniques and knowledge from all of the units in Module 2.

$$ \LARGE e^{i\pi} = -1 $$