DATAFEST Quiz-Level 1

Priya, Aria, and Ava are having a debate. According to them, which of the following best describes the principal goal of data science?

• • To collect and archive exhaustive data sets from various source systems for corporate record keeping uses.

• • To mine and analyze large amounts of data in order to uncover information that can be used for operational improvements and business gains.

• • To collect and prepare data for use as part of analytics applications.

Answer explanation

The main purpose of a data science initiative is to analyze data in ways that provide useful information to a company. That can include a mix of structured, unstructured and semistructured data, typically in large amounts that make it difficult to find meaning in the data without using advanced analytics approaches. Some common data science applications in businesses include anomaly detection to aid in fraud detection and cybersecurity efforts; pattern recognition for analyzing customer purchases, stock trading and other use cases; and predictive modeling of customer behavior, market trends and financial risks.

Ethan, Henry, and Charlotte are on a quest to become data scientists! What is the first step they should take in their data science journey?

• • Collecting data and preparing it for analysis.

• • Experimenting with and tuning different analytical models.

• Defining an analytical hypothesis that could provide business value.

Answer explanation

Developing a machine learning or statistical model that delivers useful information starts with understanding business needs and objectives and identifying a business-related hypothesis to test. That's the case even when data scientists aren't given specific business questions to answer. Ensuing steps in the data science process include data collection and preparation, trial runs of multiple analytical models, implementation of the best model to analyze the data and presentation of the results to business executives and operational workers.

Evelyn, David, and James are having a debate. They are trying to figure out the primary difference between a data scientist and a data engineer. Can you help them?

• • David thinks that a data engineer collects and prepares data, and a data scientist then analyzes it.

• • Evelyn believes that a data engineer analyzes data after a data scientist collects and prepares it.

• James suggests that a data engineer builds data pipelines and helps prepare data, while a data scientist is responsible for data collection, preparation and analysis.

Answer explanation

Data scientists take the lead in identifying, preparing and analyzing relevant data. But they often have assistance from data engineers, who help streamline analytics projects by handling much of the upfront work required to get data into the hands of data scientists. For example, they build data pipelines to consolidate data from different source systems, then aid in integrating, cleansing and preparing the data for analysis; they might also help with deployment and maintenance of analytical models. The other members of a data science team often include data analysts, machine learning engineers and data architects, who also aid in the analytics process

Imagine Grace, Daniel, and Isla are in a heated debate! They're discussing whether a successful data scientist needs a mix of technical skills, nontechnical ones, and the right personality traits. Who do you think is right?

• They all are right

• None of them are right

Answer explanation

In general, data scientists require a diverse set of skills and characteristics. That includes knowledge of statistics and mathematics, as well as technical skills in programming, predictive modeling, machine learning and deep learning, AI, data preparation and other areas. The best ones also have various soft skills and traits, such as curiosity, problem-solving and critical thinking abilities, and communication and collaboration skills. Business knowledge is important, too, to ensure that data science initiatives produce accurate and meaningful results.

Tags

CCSS.RL.11-12.3

CCSS.RL.6.6

CCSS.RL.7.6

CCSS.RL.8.6

CCSS.RL.9-10.3

Imagine William, Jackson, and Mia are in a coding competition. They are asked to use the most popular programming languages among data scientists. Which languages should they choose?

• • C and C++

• Python, R and SQL

• • Java and JavaScript

Answer explanation

Python is the programming language that's used most widely by data scientists, followed by SQL and R, according to an annual survey on data science and machine learning conducted by Google subsidiary Kaggle. Julia is a newer language that's also among the top tools and technologies for data scientists. Reflecting Python's status as the leading language, the list includes a variety of Python frameworks and libraries that can be used to support analytics applications and data visualization.

Imagine Aria, Isla, and James are data scientists. Which of the following analytical and statistical techniques do you think they would commonly use in their work?

• • Classification

• • Regression

• • Clustering

• All of these

Answer explanation

Various classification, regression and clustering methods are key data science techniques used in analytics applications to identify relationships between different data elements. Examples include decision trees and naive Bayes classifiers for classifying data into categories; linear regression and multivariate regression; and k-means clustering and hierarchical clustering. Association analysis is another technique, akin to clustering, that's done to find association rules between related data points.

Tags

CCSS.HSS.IC.B.3

Imagine Harper, Rohan, and Oliver are in a heated debate about machine learning. They are trying to figure out the primary difference between supervised and unsupervised learning. Can you help them out?

• Supervised learning involves data that has been labeled and classified, while unsupervised learning data is unlabeled and unclassified.

• • Supervised learning is monitored closely by data scientists, while they don't play a role in unsupervised learning.

• • Supervised learning is only used for image recognition, while unsupervised learning can be used for various analytics applications.

Answer explanation

Supervised and unsupervised learning are common machine learning methods. In supervised learning, labeled and classified training data is used to teach a machine learning model to produce a particular output. The goal is to enable the model to identify specified relationships and patterns in larger data sets. Conversely, in unsupervised learning, a data scientist runs an algorithm against training data that's unlabeled and unclassified. Because the desired output is undetermined, the machine learning model groups data together and identifies similarities and patterns on its own. Semi-supervised learning is a hybrid approach in which some of the training data is labeled.

Evelyn, Henry, and Grace are having a debate. Evelyn says that bias is a common problem in data science applications. Henry disagrees. Who do you think is correct?

• Evelyn

• Henry

Answer explanation

To ensure that analytics results are accurate, data scientists must be able to identify and eliminate various forms of bias both before and during the data analysis process. That can include underlying issues in the data sets being analyzed, as well as biases that inadvertently get built into machine learning algorithms and analytical models. Some examples are data that doesn't fully represent different groups of people and algorithms that propagate a current state, such as hiring analytics tools that show a preference toward men. Cognitive, confirmation and outlier biases can also skew results.

Tags

CCSS.RL.11-12.3

CCSS.RL.6.3

CCSS.RL.7.6

CCSS.RL.8.6

CCSS.RL.9-10.3

Aiden, Nora, and Avery are data scientists working on a big project. Why would they find data sampling useful?

• • It allows them to analyze data sets in small, manageable batches, reducing their use of system resources.

• • It reduces the amount of data storage space that's required for their data science applications.

• It enables them to use a representative subset of data to build accurate analytical models more quickly.

Answer explanation

If analyzing a full data set is difficult or would take too much time, data sampling can reduce it to a smaller, more manageable size. Various methods can be used to create representative samples, depending on the particular data set and planned analytics application. Done properly, data sampling produces accurate results more efficiently than would be possible otherwise. To avoid sampling errors, though, data scientists need to make sure that samples accurately represent the data as a whole.

Tags

CCSS.7.SP.A.1

CCSS.HSS.IC.A.1

Imagine Michael, Priya, and David are data scientists. What is a common saying about their work routine?

• They spend 80% of their time finding and preparing data and 20% analyzing it.

• • They spend 80% of their time analyzing data and 20% finding and preparing it.

• • They spend 80% of their time on failed analytics projects and 20% doing useful work.

Answer explanation

Collecting relevant data and preparing it for analysis are time-consuming chores for most data scientists. Data sets from multiple source systems often must be combined and consolidated, and the data preparation process includes various steps, from data profiling and cleansing measures that fix data quality problems to data transformation, enrichment and validation. This work does take a lot of time, but it's a crucial prerequisite to accurate data science applications.