DSS110 DATA SCIENCE FLASHCARDS

Which three disciplines combine to form the interdisciplinary field of Data Science?

Domain Expertise, Mathematics & Statistics, and Computing Skills.

What is the primary objective of using scientific methods and algorithms in Data Science?

To extract knowledge and insights from structured and unstructured data.

In the 'Data to Decisions' flow, how is 'Data' defined?

Raw facts and figures.

Within the context of data transformation, what characterizes 'Information'?

Organised and meaningful data.

What level of understanding is reached when 'Knowledge' is derived from information?

Understanding derived from information.

In the sequence from data to decisions, what is the role of an 'Insight'?

To identify patterns that lead to action.

What is the final stage of the 'Data to Decisions' process that leads to better results?

Decision (informed choices).

Which era of data evolution (1960s-1970s) focused primarily on hypothesis testing and traditional statistics?

Statistical Analysis.

The era of 'Data Mining' (1980s-1990s) shifted focus towards which specific activity?

Discovering patterns in data with a focus on algorithms.

Which period in the evolution of data science is associated with the rise of the '4 V's' (Volume, Variety, Velocity, Veracity)?

Big Data (2000s-2010s).

How does modern Data Science (2010s-Present) differ from previous eras in terms of value?

It focuses on end-to-end value creation by integrating statistics, ML, computing, and domain knowledge.

List the five 'V's' that characterize Big Data.

Volume, Velocity, Variety, Veracity, and Value.

In Big Data, the term _____ refers to the massive amount of data generated.

Volume

In Big Data, the term _____ refers to the speed at which new data is generated and processed.

Velocity

In Big Data, the term _____ refers to the different types of data (structured, numeric, unstructured).

Variety

In Big Data, the term _____ refers to the messiness or trustworthiness of the data.

Veracity

What are two common examples of Big Data 'Distributed Storage' systems?

Hadoop HDFS and Cloud Storage.

Name two major 'Processing Engines' used in Big Data infrastructure.

MapReduce and Spark.

Identify three examples of NoSQL databases used for Big Data.

MongoDB, Cassandra, and HBase.

Which tools are typically used for 'Data Ingestion' in a Big Data pipeline?

Kafka and Flume.

What are the three primary cloud platforms used for Big Data infrastructure?

AWS, Azure, and GCP.

Term: Artificial Intelligence

Definition: Machines that can perform tasks that typically require human intelligence.

Term: Machine Learning

Definition: A subset of AI that allows machines to learn from data.

Term: Deep Learning

Definition: A subset of ML using neural networks with many layers.

How many steps are identified in the provided Data Science Lifecycle overview?

Eight steps.

What is the first step in the Data Science lifecycle?

Problem Understanding.

In the 'Problem Understanding' phase, what must be determined to measure project success?

Success metrics.

What is the final task in the 'Problem Understanding' phase before moving to data collection?

Translating the business problem into a Data Science problem.

What are the three categories of data types identified in the 'Data Collection' phase?

Structured, Unstructured, and Semi-structured.

Aside from collecting and storing data, what is a critical quality concern in the 'Data Collection' phase?

Ensuring data quality.

What activity in 'Data Preparation' involves filling in or removing data that isn't present?

Cleaning missing values.

In 'Data Preparation', the process of creating new input variables from raw data is called _____.

Feature engineering

What is the primary goal of Exploratory Data Analysis (EDA)?

To understand patterns and relationships in the data.

Which lifecycle step involves visualising distributions and generating hypotheses?

Exploratory Data Analysis (EDA).

What are 'hyperparameters' in the context of the 'Modeling' phase?

Adjustable parameters that are tuned to improve model performance.

Which technique is used during the 'Modeling' phase to ensure the model generalises well to new data?

Cross-validation.

Identify four common metrics used in the 'Evaluation' phase.

Accuracy, RMSE, Precision, and Recall.

What is the main risk the 'Evaluation' phase seeks to avoid regarding model complexity?

Overfitting.

In the 'Deployment' phase, what does 'handling data drift' involve?

Managing changes in data properties over time that affect model performance.

What is the purpose of 'Monitoring performance' after a model is deployed?

To ensure the model remains accurate and effective in production.

What is the final step of the Data Science lifecycle?

Communication.

In the 'Communication' phase, what should be provided to help stakeholders make decisions?

Actionable recommendations.

According to the key takeaways, which two skills are as important as modelling in a Data Science project?

Communication and domain understanding.

How does Data Science help organisations regarding their 'customers'?

It helps organisations understand their customers.

Data Science creates business value and _____ _____.

Competitive advantage

The sequence of the first four steps in the Data Science lifecycle is: Problem Understanding, _____, Data Preparation, and EDA.

Data Collection

The sequence of the last four steps in the Data Science lifecycle is: Modeling, _____, Deployment, and Communication.

Evaluation

In the context of EDA, what does the acronym 'EDA' stand for?

Exploratory Data Analysis.

Which lifecycle step follows 'Modeling'?

Evaluation.

Which lifecycle step precedes 'Data Preparation'?

Data Collection.

What is the primary role of data in modern organisational strategy according to the source?

Data drives decisions.

Why is 'Feature Engineering' considered a part of 'Data Preparation'?

Because it involves transforming raw data into a format more suitable for modelling.

What is the specific goal of 'normalisation' in the Data Preparation phase?

To scale numeric data to a standard range to improve model training.

In 'Data Collection', what is the difference between structured and unstructured data?

Structured data is highly organised (e.g. databases), while unstructured data lacks a pre-defined format (e.g. text, images).