Big Questions for Social Media Big Data Ignacio Penas 27/03/2020 - - PowerPoint PPT Presentation

Big Questions for Social Media Big Data

Ignacio Penas 27/03/2020

Big data for human behaviour study

Scientific method Social study method

u

Model organism problem

u

Dependent variable selection

u

Vague sampling

u

Single platform studies

u

…

What can we infer from social big data?

Online hidden meanings Unmodalable actions

Methodological Considerations

Key points for Social Behavior analysis with big data

Twitter as the Field’s Drosophilia Melanogaster

Model organisms Why Twitter?

u

Describe why is twitter so widely used

Features analyses

I.

Hashtag analyses

I.

Affected by user activity

II.

Dependent variable selection

• III. Selection effects
• IV. User choices

The Missing Denominator

u

Data normalization

u

Non-registered visits

u

Entire population vs Sampling group

u

Platform information segregation

u

Curated information in intermediaries to provide validity to the data

Ecology for the Platform

#rt #rt #seen #users

Twitter Broadcast Media Cell-phone Face2Face FB

Inferences and Interpretations

Examination of inference concept for socual media big data interpretation

Understanding the data

u

Retweet interpretations

Affirmation Denunciation Approval Rejection Sarcarsm

Invisible Engagement

u Subtweeting u Screen capture discusions u Hate-linking

Methodological limits

u

Method importing difficulty

u

Epidemiological analyses connected to social media networks

u

Physical proximity

Field effects

u

Non-networks interactions

u

Sudden big impacts vs slow (or not so slow) spread of virus

u

Tunisian revolution effect over the Egiptian

u

Real meaning depends on context (space and time)

Conclusions

I.

Target non-social dependent variables

II.

Quialitative pull-outs

• III. Baseline panels
• IV. Industry Outreach

V.

Multi-disciplinary teams

• VI. Methodological awareness

Thank you!