SI unit 2

How does Field Cady define data science?

A process-driven discipline: framing problems → understanding messy data → extracting features → modeling → presenting/deploying results. Distinct from software engineering because it is iterative and exploratory. At its core, it is turning observations into knowledge.

Source: The Data Science Road Map – Field Cady

How does Max Shron describe data analysis vs. statistics?

Data analysis is inductive, predictive, and designed to generate new knowledge from messy real-world data. Statistics is deductive and largely descriptive. Data analysis transforms insights into actionable meaning for decisions.

Source: Thinking with Data – Max Shron

What is big data according to Tiell & O’Connor?

More than “large datasets.” Big data involves volume (scale), velocity (real-time generation), and variety (different structures like text, social relationships, sensors). It creates risks of bias amplification, consent violations, and inequity.

Source: Building Digital Trust – Tiell & O’Connor

What historical example shows that “big data” is not new?

The 1890 U.S. Census used Hollerith punch cards (early big data processing).

Source: 50 Years of Data Science – David Donoho

What are the two main types of deliverables data scientists produce?

(1) For humans – reports, visualizations, presentations for decision-making.

(2) For machines – production code/models for automation, batch analytics, or real-time systems.

Source: The Data Science Road Map – Field Cady

What key skills should a data scientist have?

Technical: programming, quantitative analysis, feature engineering, modeling, visualization.

Soft: scoping problems, communication, storytelling, teamwork, product intuition.

Source: The Data Science Road Map – Field Cady; Lecture notes

What is the “snarky definition” of a data scientist from consensus curricula?

“A data scientist is better at stats than any software engineer, and better at software engineering than any statistician.”

Source: 50 Years of Data Science – David Donoho

How did John Tukey (1962) shape the history of data science?

He argued that “data analysis” should be its own science, with workflows, visualization, and empirical validation.

Source: 50 Years of Data Science – David Donoho

How did Leo Breiman’s “Two Cultures” paper (2001) influence data science?

He distinguished between generative modeling (traditional statistics) and predictive modeling (machine learning), arguing that statistics ignored prediction.

Source: 50 Years of Data Science – David Donoho

What is Max Shron’s main warning about starting data projects?

Don’t start with a dataset or tools. First scope the project using CoNVO: Context, Need, Vision, Outcome.

Source: Thinking with Data – Max Shron

What are the key stages of the data science roadmap?

Frame the problem → Understand the data → Extract features → Model → Present results or Deploy code.

Source: The Data Science Road Map – Field Cady

Why is iteration essential in data science?

Unknowns in data/features/models require constant back-and-forth. Early results and refinements save time and help avoid sunk costs.

Source: The Data Science Road Map – Field Cady

What tools does Shron suggest for aligning teams early?

Mockups (visual previews of results) and argument sketches (reasoning outlines). These clarify expectations and sharpen the project scope.

Source: Thinking with Data – Max Shron

Why is communication central in data science?

Because results are for mixed audiences (business, technical, policy). Clear storytelling is needed to overcome bias and intuition.

Source: The Data Science Road Map – Field Cady; Thinking with Data – Max Shron

What do Tiell & O’Connor mean by “digital trust”?

The belief that an organization is safe, transparent, reliable, and truthful in its data practices. It is hard to build but easy to lose.

Source: Building Digital Trust – Tiell & O’Connor

Why do they argue ethics now extend “beyond cybersecurity”?

Risks today include bias, consent violations, and unfair outcomes, not just technical breaches. Ethics must be integrated across the data supply chain.

Source: Building Digital Trust – Tiell & O’Connor

What principles are included in a new code of data ethics?

• Respect people behind the data

• Consider downstream uses

• Ensure transparency & accountability

• Treat compliance as the floor, not the ceiling

Source: Building Digital Trust – Tiell & O’Connor

What example illustrates harm from unethical data use?

A dating app amplified racial/ethnic bias in its algorithm to drive engagement, scaling social harm.

Source: Building Digital Trust – Tiell & O’Connor

What is an algorithm?

A set of instructions or rules to follow to complete a task. Examples: recipes, GPS routes, spam filters.

Source: Notes on Machine Learning, AI, and Algorithms (SI110)

What is machine learning (ML)?

A subset of AI where algorithms learn from data to make predictions/classifications, improving automatically with experience.

Source: Notes on Machine Learning, AI, and Algorithms (SI110); Machine Learning for Everyone

What is artificial intelligence (AI)?

The broad field of replicating human-like behaviors (planning, learning, reasoning, perception). Includes Narrow AI, AGI, and potential superintelligence.

Source: Notes on Machine Learning, AI, and Algorithms (SI110)

What are the benefits of AI/ML?

Efficiency, pattern recognition, automation, personalization. Examples: fraud detection, healthcare diagnostics, personalized learning, autonomous driving.

Source: Notes on Machine Learning, AI, and Algorithms (SI110)

What are the main concerns of AI/ML?

Bias (racial, gender, class), privacy violations, job loss, deepfakes/misinformation, weaponization, lack of accountability, surveillance.

Source: Notes on Machine Learning, AI, and Algorithms (SI110); Biases in AI Systems – Srinivasan & Chander

How do Srinivasan & Chander classify AI bias?

Bias can occur at all stages: data creation (sampling/labeling), problem formulation, analysis (proxies, confounding), and evaluation/validation.

Source: Biases in AI Systems – Srinivasan & Chander (2021)

How does Safiya Noble argue algorithms can reproduce inequality?

Search engines are ad-driven, not neutral. They amplify racism and sexism (e.g., pornification of “Black girls” search results; radicalization pathways).

Source: Algorithms of Oppression – Safiya Noble

What is information visualization?

Computer-supported visual representations of abstract data to amplify cognition (e.g., dashboards, scatterplots).

Source: Information Visualization Lecture Notes

What are the key functions of visualization?

To record, analyze, aid memory, communicate/persuade.

Source: Information Visualization Lecture Notes

What is the difference between information visualization and infographics?

Visualizations are for analysis/interaction; infographics are for storytelling and communication (often simplified).

Source: Information Visualization Lecture Notes

Why might visualization matter more than statistics alone?

Example: Anscombe’s Quartet – datasets look identical statistically but appear very different when graphed.

Source: Information Visualization Lecture Notes

What are “aesthetics” in visualization?

Visual properties like position, size, shape, color that encode data values.

Source: Visualizing Data: Mapping Data onto Aesthetics – SI Textbook

What is prompt engineering?

The process of refining inputs to generative AI so outputs are useful, accurate, and aligned.

Source: AI Made Simple (Kapur)

What are some strategies for effective prompting?

Be specific and clear; provide context; use step-by-step instructions; rephrase when needed; avoid bias; set constraints.

Source: AI Made Simple (Kapur)

What is prompt chaining?

Iteratively refining prompts (using follow-ups) to guide the AI toward a desired result.

Source: AI Made Simple (Kapur)

What is the equity issue in data visualization?

Most visualizations assume able-bodied users, excluding people with visual, cognitive, or motor disabilities — affecting >1B people globally.

Source: Inclusive Data Visualization for People with Disabilities: A Call to Action – Marriott et al. (2021)

What are barriers for users with visual impairments?

Limited alt text, lack of tactile/sonification tools; current solutions (like SAS Graphics Accelerator) often incomplete or costly.

Source: Inclusive Data Visualization… – Marriott et al.

What are barriers for users with cognitive/learning disabilities?

Struggles with abstraction and symbolic conventions; suggested fixes include chunking, coupling visuals with text, and guided exploration.

Source: Inclusive Data Visualization… – Marriott et al.

What are barriers for users with motor disabilities?

Difficulty with interactive features like zoom or lassoing; alternatives include eye tracking, speech input, or custom timing controls.

Source: Inclusive Data Visualization… – Marriott et al.

What is the main call to action in Marriott et al.’s article?

Build multimodal, evidence-based, inclusive visualization tools; involve disability communities; treat accessibility as central to design.

Source: Inclusive Data Visualization… – Marriott et al. (2021)

What benefits of social media are highlighted in Social Media for Public Health (Jafar et al., 2023)?

Provides direct access to health info, peer support, and community; enhances patient–provider connection; enables rapid alerts in crises; reduces isolation; and fosters positive mental health interactions.

Source: Jafar et al., Social Media for Public Health (2023)

What drawbacks of social media are identified in Jafar et al. (2023)?

Amplifies misinformation, polarizes users, harms youth mental health (anxiety, depression, body image issues), fuels overuse, promotes unsafe self-diagnosis, and exposes teens to cyberbullying and harmful challenges.

Source: Jafar et al., Social Media for Public Health (2023)

What does Khalaf et al. (2023) find about social media and adolescent mental health?

Dual role: builds social connection, access to resources, and identity expression; but correlates with depression, anxiety, sleep disruption, body dissatisfaction, and cyberbullying. Effects depend on individual vulnerabilities.

Source: Khalaf et al., Impact of Social Media on Mental Health of Adolescents and Young Adults (2023)

According to Owens & Lenhart (2020), what myths distort how we think about social media’s impact?

Myths like “social media is addictive,” “tech companies can fix harms with tech,” and “less screen time = healthier” oversimplify experiences and ignore structural inequities.

Source: Owens & Lenhart, Good Intentions, Bad Inventions (2020)

How does Sherry Turkle describe the benefits of social media for teens?

Provides connection, intimacy, collaborative identity, and new ways to manage fear/loneliness; allows identity play via avatars and profiles.

Source: Turkle, Adolescents & Social Media (lecture/reading notes)

What are the drawbacks Turkle emphasizes?

Fragile identities dependent on validation, anxiety from constant texting, risks like texting while driving, reduced privacy, parental surveillance, and exhaustion from self-curation.

Source: Turkle, Adolescents & Social Media

Define social media.

Platforms that enable user-generated content, sharing, and direct user-to-user interaction (e.g., Facebook, TikTok, Reddit). Also serve as stages for identity performance and self-presentation.

Source: Lampe lecture; Turkle, Adolescents & Social Media

Define social capital in the context of social media.

The resources and benefits individuals gain through online networks, such as support, validation, and status. Includes both strong-tie support and weak-tie informational access.

Source: Lampe lecture; Turkle, Adolescents & Social Media

Define the networked self.

Identity shaped through constant digital connectivity and peer validation. Adolescents co-create selfhood in dialogue with others, often fragmented across multiple profiles.

Source: Papacharissi (concept), reinforced in Turkle readings

How does Lampe frame the “negative narrative” about social media?

Critics claim social media replaces authentic intimacy with shallow ties, creates envy/anxiety, and fosters harassment. These echo historical moral panics (jazz, telephones, writing).

Source: Lampe, Social Media Is Good for You

What benefits of social media does Lampe highlight?

Builds social capital, strengthens weak ties, enables grassroots organizing, lowers mobilization costs, fosters lightweight but meaningful “social grooming,” and empowers users as their own media.

Source: Lampe, Social Media Is Good for You

What is Lampe’s overall stance in the debate?

Social media is not inherently good or bad—it’s what we make of it. With media literacy, intentional design, and user responsibility, it can be a force for positive connection.

Source: Lampe, Social Media Is Good for You

How do Khalaf et al. (2023) echo Lampe’s framing of the debate?

They emphasize a balanced view: social media can build connection and resilience but also creates risks. Outcomes depend on context, use, and individual vulnerabilities—not a simple “good vs bad.”

Source: Khalaf et al., Impact of Social Media on Mental Health

How does Turkle’s perspective connect to Lampe’s debate framing?

Like Lampe, she rejects binary thinking: technology affords both benefits (connection, visibility, identity play) and drawbacks (anxiety, fragile selves, surveillance). The impact depends on cultural context and use.

Source: Turkle, Adolescents & Social Media