How Social Data Shapes Your Relationships, Emotions, and Reality
Series: Seeing the Social Program (1/4)
The System Behind the Screen
You send messages. You scroll. You respond.
It feels like a series of small, independent actions. But across modern platforms, these actions generate structured data that can be collected, modeled, and interpreted at scale.
How we communicate, when and for how long we engage, and who we interact with... when take togeter, this creates a layered system:
A model of your relationships, your responses, and the environment around you.
This post breaks that system into three parts:
- Structure — how your relationships are mapped
- Emotion — how your responses are inferred
- Environment — how groups and information flows are shaped
When you send a message, more than content is created.
There is also metadata tagged for both the sender and recipient, like time, frequency, response patterns, etc.
In 2018, reporting showed that Facebook collected call logs and SMS metadata from some Android users through granted permissions.¹ While message content was not collected in those cases, research demonstrates that metadata alone can be highly revealing.
A study from Stanford University found that telephone metadata can be used to infer sensitive personal attributes and reconstruct relationship patterns with notable accuracy.²
What This Enables
From these signals, systems can estimate:
- Relationship strength
- Communication priority
- Social structure over time
This creates a functional map of:
Who matters most in your network
Your messages may feel private, but the structure around them can be used to:
- Model relationships
- Track changes over time
- Prioritize interactions
Your Emotions as Signals
From Behavior to Inference
Beyond structure, systems analyze behavioral signals like word patterns, interaction timing, and engagement levels. These are used in sentiment analysis and behavioral modeling.
They do not directly measure emotion. But they can infer patterns correlated with emotional response.
Evidence at Scale
In 2014, Facebook conducted a study involving approximately 689,000 users.³
By adjusting the emotional tone of users’ feeds, researchers observed measurable changes in user-generated content.
The study concluded that:
- Exposure influenced expression
- Emotional signals could propagate through networks
Optimization Through Engagement
In 2021, reporting by The Wall Street Journal showed that engagement-based ranking systems could amplify content that generated stronger reactions.⁴
These systems rely on:
- Click behavior
- Interaction patterns
- Time spent
While not labeled as emotion, these signals function as proxies for response intensity.
Key Realization
Systems don’t need to know exactly how you feel.
They only need to:
- Recognize patterns
- Reinforce responses
- Optimize for engagement
Your Environment as a System
From Individuals to Networks
At scale, individual data becomes network data.
Systems represent your social world as a graph:
- Individuals → nodes
- Relationships → connections
- Interaction strength → weights
Research in network science (including work associated with MIT and Stanford University) shows that such graphs can be used to:
- Predict information flow
- Identify influential individuals
- Model group behavior patterns
Real-World Application
During the Cambridge Analytica scandal, data from millions of Facebook users was used to:
- Build psychographic profiles
- Segment audiences
- Deliver targeted messaging
Investigations by regulators, including the UK Information Commissioner’s Office, found that data analytics techniques were applied to influence communication strategies at scale.⁵
How Influence Actually Works
At the group level, influence is rarely direct.
It operates through content selection, control of visibility, and repetitive and reinforcing delivery.
This affects:
- What conversations persist
- Which ideas gain traction
- How group norms evolve
Key Realization
You are not influenced in isolation. Your environment is shaped over time, all around you, more often than not, by other people.
The Full System
This progression is consistent:
First, your relationships are mapped.
Then, your responses are inferred.
Finally, your environment is shaped.
Why This Matters
These systems influence:
- Attention
- Perception
- Social dynamics
Not through direct control, but through structured environments that guide behavior.
Companion Intelligence — A Different Model
Companion Intelligence is designed around a different assumption:
Your data should serve you and not the business objectives of the data marketplace.
Core Principles
- Local Memory
- Private Context
- User-Controlled Recall
What This Enables
Instead of being modeled, you can ask:
- “When did I last speak with this person?”
- “What commitments did we make?”
- “How have my relationships changed?”
Without contributing to systems designed to influence behavior.
Closing
If a system maps your relationships, it can understand your world.
If it models your responses, it can anticipate your behavior.
If it shapes your environment, it doesn’t need to control you.
It only needs to guide what you see.
