Data Science, AI, and Analytics: Key Concepts and Industry Insights

One of the points made by Peters in his article describing what he learned his first year as a data scientist is that understanding the daily workflows and pain-points of stakeholders is the first step to being a successful data scientist.

True

Business strategies focus on how to execute and be successful with the organization's business model. 

True

Which of the following is the best example of an analytics application?

A dashboard application.

What is not true of the latest release of ChatGPT?

There have been few complaints about it.

Citizen data scientists possess all the skills of formally trained data scientists.

False

Which of the following is not a topic that has received attention recently? 

Perplexity being bought by OpenAI.

In organizations mature in the use of analytics, it is common to have both centralized and decentralized analytics teams (i.e., groups)

True

Which of the following came first in the evolution of decision support (i.e., business analytics)?

Decision support systems.

Which of the following companies illustrate using analytics to seize a market opportunity?

Harrahs.

A review mirror in a car is analogous to descriptive analytics.

True

The AI big tech companies (e.g., OpenAI, Google) have recently announced AI learning initiatives and programs. 

True

Preattentive attributes like color and size are automatically processed by the human brain 

True

How does a Gantt chart visualize project activities?

Using timelines for project tasks.

The big tech AI market leaders are now experiencing problems in raising funding for future initiatives. 

False

Fine-tuning a foundation model involves adding specific application data to improve its output for particular tasks.

True

Structured prompts can be a useful addition when creating a custom GPT.

True

Reinforcement Learning form Human Feedback (RLHF) is used to improve the accuracy of responses in generative AI. 

True

What did we not learn from Andi Brock's presentation?

The need for technical skills is the most important part of what she does.

A best practice for companies is to discourage the emergence of citizen data scientists because CDS are not well trained in analytics.

False

What is one of the first tasks an Analytics Director should undertake when starting a new position?

Work with senior management to develop analytics strategies that are aligned with the business strategies 

It is best to test a predictive model using the same data it was trained on.

False

There may be overfitting when a model performs well on training data but poorly on new data.

True

Which of the following statements is not true of BigML?

It allows the input data to be in a format other than rows and columns.

A limitation of BigML is that it does not show the distribution of the input variables.

False

BigML allows the user to split the data into training and test data sets, including the percentage in each set and how the sets are selected (e.g., random).

True

Which of the following is not a recent development with generative AI?

Open AI announced that ChatGPT will not be used for ecommerce

Which type of information does market basket analysis provide?

Association information

What is the IRS’ Discriminatory Index Function (DIF) score used to determine?

The best candidates for an audit

OLAP and predictive analytics are the same concept and the terms can be used interchangeably.

False

A confusion matrix is used when evaluating the predictive power of a model with a continuous dependent variable.

False

Which of the following statements is not true of recent developments in AI?

The concerns about AI becoming too powerful have subsided

Some data can be used in some analytics applications but not in others.

True

What is the main purpose of the creepiness scale?

To measure how customers react to companies' use of personal data and algorithms

What of the following was not a "takeaway" from Matt McGivern's presentation?

Companies should store personal data as long as possible in case it is needed in the future

The EU's GDPR require companies to disclose what personal data they store and to correct any inaccurate data.

True

Pre-screening bias can occur in modeling if data splitting is not performed for model building and evaluation.

False

Which of the following statements is not true of Hadoop/MapReduce

Replaces the need for a data warehouse

The large investments in data centers to support AI has driven up the stock prices of big tech companies and created fears about an "AI bubble."

True

The European Union has taken a lead in passing AI regulation with the EU AI Act.

True

Algorithmic transparency and accountability refers to companies being more public, open, and accountable in their use of personal data and algorithms 

True