Mapping Idea Spread Through Semantic Networks
Semantic networks are maps of concepts connected by their relationships.
In social media: Hashtags connected by co-occurrence (appearing together in the same post).
If these tweets exist:
"Fighting for our future! #ClimateStrike #FridaysForFuture #ClimateJustice"
"Join us tomorrow! #ClimateStrike #ActOnClimate"
"Youth leading the way #FridaysForFuture #YouthVoice"
Hashtags that appear together are connected
19 million tweets in first year. Changed global conversation about sexual harassment.
47 million tweets in May 2020 alone. Largest protest movement in US history.
Youth-led global movement. 7+ million protesters in 150+ countries.
1. Unite Communities
Create shared identity around a cause
2. Spread Ideas
Make content discoverable and viral
3. Signal Participation
"I'm part of this movement"
If hashtags drive movements, who controls which hashtags gain traction?
By mapping hashtag networks, we can see:
Gather 10,000 tweets about climate change
From each tweet, pull all hashtags
Example: "#ClimateStrike #FridaysForFuture #ActNow" → ["ClimateStrike", "FridaysForFuture", "ActNow"]
Count how often each pair of hashtags appears together
Nodes = hashtags, Edges = co-occurrence strength
Result: A map showing how ideas connect through shared language
Here's what a co-occurrence matrix looks like for climate hashtags:
| Hashtag Pair | Co-occurrence Count |
|---|---|
| #ClimateStrike + #FridaysForFuture | 1,247 |
| #ClimateStrike + #ClimateJustice | 892 |
| #ClimateStrike + #ActOnClimate | 654 |
| #FridaysForFuture + #YouthVoice | 412 |
| #ClimateChange + #FossilFuels | 389 |
Thicker lines = stronger co-occurrence
#ClimateStrike
⬍ (1,247) ⤏
#FridaysForFuture
⬍ (412) ⤏
#YouthVoice
This network shows the semantic structure of climate activism language on Twitter.
Once you have a hashtag network, analyze it using these metrics:
Which hashtags are most connected?
Example:
#ClimateStrike has 47 connections
#ActNow has 8 connections
→ #ClimateStrike is more central to the movement
Interpretation: Central hashtags = unifying language for the community
Are there distinct sub-groups?
Example:
Cluster 1: #ClimateStrike, #FridaysForFuture, #YouthVoice (Youth activists)
Cluster 2: #GreenNewDeal, #ClimatePolicy, #RenewableEnergy (Policy focus)
Interpretation: Different sub-communities with distinct language
Which hashtags connect different communities?
Example:
#ClimateJustice appears in both youth activist AND policy clusters
→ It bridges the two communities
Interpretation: Bridging hashtags spread ideas between groups
Key Insight: Network structure reveals how language organizes social movements.
Critical thinking checkpoint
Analyze this scenario:
Hashtag Network: #ClimateChange
Two Clusters Found:
Bridging hashtags: #ClimateChange, #Science
What does this network reveal about competing narratives?
Which hashtags "win"?
Data:
Activist hashtags are 2.7x more central.
Does network centrality = narrative power?
Consider:
Why do some hashtags go viral while others don't?
Factors:
Who has power in hashtag virality? Users? Platforms? Media?
Design a platform that promotes bridge-building hashtags:
How would you design Twitter/X to:
Propose algorithm changes to the "Trending" feature.
Not all hashtags serve the same purpose. Research identifies three types:
Purpose: Organize, inspire participation, rally people
Characteristics:
Purpose: Signal belonging to a group/ideology
Characteristics:
Purpose: Categorize content, make discoverable
Characteristics:
Social movements strategically combine these types:
#ClimateStrike (Mobilizer) + #YouthActivist (Signifier) + #ClimateChange (Classifier)
= Call to action + identity + discoverability
Classify these 5 hashtags into: Mobilizer, Signifier, or Classifier
In a full implementation, this card would contain an interactive network graph using D3.js or Cytoscape.js.
In the live version, you could:
Upload your CSV of tweets → VineAnalyst extracts hashtags → Builds co-occurrence network → Generates interactive visualization
Critical thinking checkpoint
Research these famous hashtag hijacking examples:
NYPD asked people to share positive photos with police.
Result: People flooded it with photos of police brutality.
McDonald's wanted heartwarming customer stories.
Result: Horror stories about food poisoning and bad experiences.
How did activists/trolls repurpose these corporate hashtags?
Why is hashtag hijacking so effective?
Consider:
Corporate brands spend millions promoting hashtags, then lose control of them.
Does hashtag hijacking redistribute power from brands to users?
Is hashtag hijacking activism or harassment?
Activists used it to highlight police brutality → Holds power accountable
Trolls hijack #NewProduct to spam fake reviews → Just harassment
Where's the line between legitimate protest and malicious trolling?
Create guidelines for brands on hashtag vulnerability:
As a social media manager, how would you:
Write a 3-rule "Hashtag Safety Checklist" for brands.
Hashtag network analysis is used by:
Use case: Tracking misinformation spread
Example: ProPublica analyzed anti-vaccine hashtag networks to trace misinformation origins
Use case: Brand sentiment monitoring
Example: Nike tracked #JustDoIt network during Colin Kaepernick campaign
Use case: Tracking movement evolution
Example: Researchers mapped #BlackLivesMatter network to see policy focus vs. direct action split
Use case: Detecting coordinated inauthentic behavior
Example: Twitter used hashtag networks to identify state-sponsored influence operations
Common thread: Hashtag networks reveal who is connected to whom through shared language.
✅ Semantic Networks
Ideas connected by co-occurrence. Hashtags that appear together form networks.
✅ Building Hashtag Networks
Extract hashtags → Build co-occurrence matrix → Create network graph → Analyze centrality, clustering, bridges
✅ Three Hashtag Types
✅ Real-World Applications
Track social movements, brand sentiment, misinformation spread, coordinated campaigns
⚠️ Who controls hashtag narratives?
Network centrality = power. Central hashtags shape the conversation.
⚠️ Hashtag hijacking
Brands lose control. Users redistribute power. When is it activism vs harassment?
The bigger picture: Language structures social movements. Hashtag networks show us HOW ideas spread.