PSOSM 2021 Live Lecture 1

Course Focus

The lecture introduces a course on privacy and security in online social media. Rather than surveying social media generally, it uses social-network and security concepts to study issues such as:

• How personal information is exposed or inferred from posts, images, and interactions.
• How to identify fake accounts, automated users, and misinformation.
• How social platforms can be used—or misused—to influence people and events.

Social media now supports a wide range of everyday activities, including communication, entertainment, professional networking, gaming, location sharing, and business.

Privacy and Security

Privacy concerns control over observation and information—for example, choosing who can see a post, avoiding unwanted tracking, or preventing information from being shared beyond its intended audience.

Security concerns protecting information as it is stored or transmitted, including preserving confidentiality and preventing unauthorized alteration or loss.

The lecture distinguishes between the two: information can be securely protected while still being used in ways that raise privacy concerns.

Social Media Platforms and Data

Different platforms are associated with different kinds of content or interaction, including images, video, short text, professional contacts, location, dating, blogging, anonymous posting, and ephemeral content that disappears after a period.

A platform can count as social media even if it is not a conventional social-networking site. Online games, communication tools, and collaborative sites such as Wikipedia can also involve social connections, content creation, and interaction.

The growth of social media is described through the five Vs:

• Volume: The quantity of data produced.
• Velocity: How quickly data is generated.
• Variety: The diversity of platforms and content types.
• Veracity: The difficulty of verifying whether content or accounts are genuine.
• Value: The insights, predictions, recommendations, or advertising decisions that can be derived from data.

Social-Network Concepts

Six Degrees of Separation

In a social network, two people may be connected through a relatively short chain of acquaintances. The lecture presents six as a commonly discussed average and notes that distances on platforms such as Facebook have been reported as smaller.

Strong and Weak Ties

• Strong ties are close, frequent relationships, such as those with close friends or family.
• Weak ties are looser connections or acquaintances.

Weak ties can provide access to unfamiliar ideas, information, and opportunities that may not circulate within someone’s close-knit group.

Questions Social-Media Analysis Can Address

Social-media data can help researchers examine:

• Who is discussing an event.
• When discussion occurs.
• Where posts originate, when location information is available.
• What people are discussing.

It is harder to establish why people post or exactly how they are behaving. Some technical details, such as the device used, may be observable, but they do not fully explain a user’s motives.

Potential Uses and Risks

Social platforms can support crisis response, coordination, and public discussion. Examples raised in the lecture include the Hudson River airplane landing, protests following the 2012 Delhi gang rape, and the Arab Spring.

The same data can be used to study or attempt to predict events such as protests, market changes, or health-related trends. Some analyses are retrospective; having useful data and analysis before an event is the difficult part.

Social media can also be used for recruitment into harmful organizations, and users’ likes, shares, and posts may have consequences later. The lecture raises the possibility of profiling people based on their activity and using those profiles to target messages or advertisements.

It also introduces shadow profiles: information about people who have not joined a platform may still be collected or assembled, so a profile is ready if they later sign up.

Hashtags group posts around a topic, making it easier to find and collect relevant content for analysis.

Suggested Analysis Approach

A practical workflow for studying a social-media topic might include:

1. Choose an event, issue, or community to investigate—for example, traffic, an election, a protest, or a sports match.
2. Identify relevant keywords or hashtags that can help locate related posts.
3. Collect social-media data associated with those terms, subject to the platform and available data.
4. Examine timing and activity patterns, such as sudden increases in posts or discussion.
5. Analyze content, potentially using natural-language processing (NLP) to identify topics or sentiment.
6. Analyze connections between users, using network analysis to study interactions and prominent accounts.
7. Interpret findings cautiously. A post’s timing, location, or network connections may be observable, but those signals do not necessarily reveal a person’s motives or prove that a prediction is correct.

Possible project ideas include examining discussion around an event, comparing activity across platforms, studying interactions among cricket players or commentators, detecting suspicious or automated accounts, and investigating how online content may affect users.

Course Participation and Learning Goals

Students are encouraged to use the discussion forum and mailing list, attend office hours, and ask teaching assistants for help with labs.

Planned practical work includes collecting and analyzing social-media data, including through NLP and network analysis. The instructor encourages interested students to go beyond the required coursework and explore projects or solutions based on social-media data.

Speakers and Sources Mentioned

• PK: The instructor, identified as a faculty member at IIIT Delhi.
• Neha, Vinash, and Prashant: PhD students identified as assisting with course discussions.
• “It Cool”: Mentioned as helping with lab content; the name is unclear in the auto-generated subtitles.
• The Great Hack: Documentary recommended in connection with Cambridge Analytica and political profiling.
• The Social Dilemma: Documentary recommended in connection with behavioral manipulation and social-media business models.
• Mark Zuckerberg’s U.S. Congress hearing: A video of the hearing is recommended; no additional speaker is clearly identified in the subtitles.
• Minority Report: Film mentioned as an analogy for predicting events before they happen.
• The instructor’s Parler-related report: Described as comparing Parler and Twitter during the Capitol events; no report title is given.
• Students and class participants: Contribute comments and responses but are not individually identifiable as speakers in the subtitles.