Datafication¶
Datafication renders social and institutional practices as processable digital traces, enabling tracking, comparison, and action under particular encoding and governance choices.
Core Idea¶
Datafication is the socio-technical process of translating social practices, interactions, or institutional processes into digital traces that can be collected, aggregated, and processed to make those practices trackable, comparable, or actionable. The word names a transformation of the activity's visibility: a platform's “like” or “follow” becomes a structured relation, and a school's learning activity becomes an event or record that can be displayed or analyzed. The trace is a selective account of the practice, shaped by the system that records it.[1][2]
The identity does not require that the activity was wholly unrecorded before. Nor does it require every trace to be captured passively, every record to be a scalar number, or every use to be commercial or secondary to the original activity. Those may occur in particular settings. What must remain is a practice-to-data mapping and a processable informational layer that permits tracking, comparison, profiling, or other action on the represented practice. Van Dijck analyzes this transformation in social media; Pangrazio, Selwyn, and Cumbo study it in three Australian schools.[1][2]
The data layer can help institutions notice patterns, but it is not a transparent copy of life. Features, defaults, categories, and access rights determine what becomes visible. Van Dijck criticizes an accompanying belief that large platform datasets directly and objectively reveal all social behavior; that belief is dataism, a related ideology rather than the datafication process itself.[1]
Structural Signature¶
- Practice or institutional process — target. An interaction, routine, learning activity, or social relation is the subject of the new data account. It may already have other records; datafication adds a particular digital rendering.[1][2]
- Digital mediation and encoding — transformation. Platform features, learning systems, sensors, forms, or software rules select events and categories and encode them as data. The exact instrument varies; passive by-product capture is a variant, not a requirement.[1][2]
- Processable trace — new informational layer. A retained event, relation, category, count, or metadata field can be stored, linked, aggregated, or compared. Its format allows computation while leaving out aspects of the original practice.[1][2]
- Interpretation and action — consequence channel. Some actor can read or process the trace to track, profile, compare, predict, or decide. No single downstream use is universal, and the existence of a dashboard alone does not prove a particular intervention occurred.[1][2]
- Governance and representational limit — diagnostic. Who designs the categories, accesses the data, and interprets the output affects the meaning and consequences of the trace. These questions do not predetermine that one party always controls or benefits from reuse.[1][2]
Remove the mapping from practice into processable digital data and the case becomes ordinary activity, not datafication. Remove every possible route to interpreting the trace and the account loses the tracking or analysis relation that makes it useful. Remove a particular predictive or commercial use, however, and the process may remain.
What It Is Not¶
Digitization re-encodes an existing representation, such as a paper page into an image. That act alone does not establish a new data account of the behavior that produced the page. Datafication may coexist with digitization, but its test is the mapping of a practice or process into new analyzable traces. A school can digitize a timetable and also datafy student interactions with its learning system; these are distinct operations.[2]
Informating is the live catalog's narrower organizational concept: computer-mediated work simultaneously generates information about that work. It overlaps with datafication when a work system yields usable traces, but the “automate and informate” duality is not a condition on social-platform relations or every student learning trace. Informatization concerns wider structural change toward information-intensive institutions or societies, rather than each individual rendering of practice as data.
Dataism is not a more advanced stage of datafication. Van Dijck uses it for trust in objective quantification and broad behavioral tracking. A researcher can study datafication while rejecting that trust. Likewise, surveillance, commercial targeting, and asymmetric control are possible consequences or governing arrangements, not definitional outcomes of every captured trace.[1]
Scope of Application¶
In platformed social life, datafication can turn friending, liking, following, communication, and expressions of preference into records with counts, ties, or metadata. Van Dijck describes how social interactions become algorithmic relations and how aggregated metadata can be mined and repurposed. Her article also warns that platform mediation and user selection make these traces poor candidates for an unqualified “whole population” picture.[1]
In education, Pangrazio and colleagues describe digital representations of learning practices and institutional processes used to track, profile, or predict behavior and learning. Their fieldwork covered three Victorian secondary schools and 62 students aged 14–16. The schools had learning management systems, Google Classroom, laptop and WiFi monitoring, analytics, dashboards, and reporting tools. The paper studies students' accounts of these environments; it is not a technical audit proving every possible data flow in each system.[2]
The concept can be used in other settings when the same practice-to-digital-trace relation is actually shown. A bare mention of “big data,” a digital document, or a database does not suffice. The claim should identify the activity, encoding rule, trace, processing, and feasible or observed use. The sources here establish the social-platform and three-school settings; they do not independently establish the seed's HR recruiter-ranking example or a universal claim about all institutions.
Clarity¶
Datafication distinguishes an activity from its data proxy. A like is an action within a platform interface; a count or network edge records selected features of it. The proxy may be useful for a question about platform activity while being weak evidence for a broader claim about friendship, interest, or sentiment. Van Dijck's criticism of apparently neutral platform data makes this distinction consequential rather than merely terminological.[1]
It also separates capture, processing, interpretation, and use. A school system's capacity to log student activity does not prove those logs entered a profile or that a particular prediction was made. Students in the field study described some digital records and monitoring and speculated that activity traces might augment profiles. Their speculation is evidence about perceived data practices, not a system-level trace of all profile inputs.[2]
Finally, the process and its evaluation are different. Naming datafication does not itself decide whether a use is effective, fair, representative, or harmful. Those conclusions require the local evidence and governing standards.
Manages Complexity¶
A platform or school can generate many kinds of records. The abstraction reduces this complexity to five questions: What practice is rendered? How is it encoded? What trace persists? Who can interpret or act on it? What features and perspectives are omitted? That small map makes otherwise scattered data systems comparable without pretending their purposes or power relations are identical.
This compression prevents a common shortcut: treating a large dataset as though it were a direct census of the world it represents. Platform users, interface prompts, recording rules, and categories affect what appears. A school dashboard likewise arranges selected activities into institutionally legible forms; the dashboard does not contain the full experience of learning. These are analytical limits of the mapping, not a claim that every dataset is useless.[1][2]
Abstract Reasoning¶
When evaluating a datafication claim, start with the target practice and follow one event through the system. Which action is registered, under which category, with which time and actor fields? Can records be linked across events or people? What comparisons or inferences become possible, and who can act on them? Then test the representation's fidelity: which aspects of the practice were never captured, and which apparent patterns might be artifacts of the interface or sample?[1][2]
Keep observed effects separate from possible ones. Van Dijck gives examples of platform metadata's availability and critiques assumptions about representativeness. Pangrazio and colleagues provide bounded qualitative evidence about students' experiences in three schools. Neither licenses a general assertion that every captured record is resold, every school predicts student futures, or every affected person has no say in interpretation.[1][2]
Knowledge Transfer¶
The analytical procedure transfers literally between platform and educational datafication: identify target practice, encoding, trace, processing, and interpretation. The substantive findings do not transfer automatically. A social-media relationship metric and a school's account of student work differ in purpose, actors, consent setting, and evidence for actual use. An analyst must remap those roles and recheck the record's omissions in each institution.[1][2]
The live Representation Prime describes the necessary target-to-medium mapping in either case. Datafication adds digital mediation of lived or institutional practice and the resulting processable trace. A proposed wider Prime about selective traces across non-digital practices would need separate evidence and admission; the two settings here establish a domain-specific socio-technical process.
Examples¶
Social-platform relation coding¶
Van Dijck describes how friending and liking become algorithmic relations and how communication metadata become resources for analysis. The target practice is social interaction or expressed preference. A platform feature supplies the encoding rule; the resulting ties, counts, and logs are the processable traces. Platform operators and other authorized or receiving parties may aggregate them for tracking or interpretation. The governance and fidelity check asks who can access the records and whether platform participation, interface design, or missing context makes an inference about wider social behavior unwarranted.[1]
Mapped back: social action → feature-defined digital event or relation → aggregable trace → possible tracking/analysis → source-specific access and representativeness questions. Commercial repurposing is documented in the article's platform context; it is not a required outcome of every instance.
Three Victorian schools¶
Pangrazio and colleagues studied schools with learning systems, Google Classroom, laptop and WiFi monitoring, analytics, and dashboards. The target practices include learning activities and school routines. Those systems define which actions and institutional facts can be encoded; digital records and displays form the processable trace. Interviewed students reported teacher displays, prompts, and monitoring, and some presumed or speculated that activity traces might augment profiles. The interpretation channel is therefore documented partly through students' accounts, not through an audited map of every system-to-profile flow. The governance check concerns student awareness, access, and ability to contest the resulting account.[2]
Mapped back: school practice → system-defined record → available digital traces and displays → reported tracking/accountability encounters → bounded student perspectives and unverified downstream data flows. The example illustrates the process without turning student speculation into an observed technical fact.
Structural Tensions¶
The blueprint identifies no single intrinsic design trade-off required for every datafication instance. In particular, a process does not cease to be datafication because its records are not sold, reused for a new purpose, or controlled asymmetrically. The sources do show setting-specific pressures: more detailed traces can make activity more legible while raising questions about omitted context, access, and surveillance. Those pressures should be analyzed in a concrete platform or school rather than treated as an invariant two-pole mechanism.[1][2]
Structural–Framed Character¶
Datafication sits toward the framed side of the spectrum. Evaluative weight: the process can be described without approving it, though its fairness and representativeness are contested. Human-practice dependence: its target is social or institutional practice and its categories reflect human decisions. Institutional origin: platforms, schools, and data-governance arrangements shape the traces and their uses. Vocabulary travel: “turning something into data” travels widely, but the admitted identity requires a digital practice-to-trace mapping rather than any measurement. Import versus recognition: an analyst recognizes datafication by finding that mapping in a new institution; calling a handwritten recollection “datafication” imports a loose analogy. Its character: a repeatable but institution-shaped socio-technical transformation, not a substrate-independent law.[1][2]
Structural Core vs. Domain Accent¶
The structural core is a selective target-to-medium mapping: features of a practice are encoded into an interpretable trace. That is why this process presupposes Representation. The domain accent is the digital capture and processing of human or institutional activity, with platform or organizational choices about categories, aggregation, access, and use. Remove those conditions and the entry loses its named datafication identity even though the broad representational relation may remain.
This entry does not clear the Prime bar. The sources give a social-platform analysis and a bounded school study, both inside digital social organization. They show a repeatable process there, not a demonstrated substrate-independent mechanism that would include analog practices without changing the identity. A future Prime claim would require unlike non-digital or non-institutional instances and its own admission gate; it is not obtained by dropping the word “digital.”[1][2]
Instantiates / Related Primes¶
This entry presupposes Representation.
The one recorded DAG relation is a strict composition/presupposes edge to Prime Representation. Every admitted datafication instance maps an activity or process into selected features of a digital medium for interpretation. Representation exists without datafication in analog maps, diagrams, and models, and datafication is a process rather than a kind of representational artifact. That is why a presupposed component fits better than subsumption.
Informating overlaps where digital work produces traces, but its work-specific automate/informate duality is not universal here. Informatization names a broad structural shift of institutions or society. Data Collection in the live catalog requires a deliberate question, protocol, and provenance; a platform exhaust trace need not satisfy that identity. A thematic relation to those entries does not justify another direct edge.
Relationships to Other Abstractions¶
Current abstraction Datafication Domain-specific
Parents (1) — more general patterns this builds on
-
Datafication presupposes Representation Prime
Datafication presupposes mapping selected features of a practice into an interpretable digital trace.Every admitted datafication process has a practice or institutional process as target, a digital trace as medium, an encoding rule that selects and omits features, and a convention for interpreting or acting on the trace. That mapping is Representation as defined by the live Prime and is necessary for the datafication process. Datafication is a process that presupposes this mapping, not a subspecies of representations. Analog maps and models instantiate Representation without datafication, making the composition relation strict.
Hierarchy path (1) — routes to 1 parentless root
- Datafication → Representation → Abstraction
Neighborhood in Abstraction Space¶
Datafication sits in a sparse region of the domain-specific corpus (96th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.
Family — Semantic Integration & Knowledge Modeling (13 abstractions)
Nearest neighbors
- Informating — 0.84
- Social Semantic Web — 0.77
- Flow process chart — 0.77
- Transformed Social Interaction — 0.77
- Gradual release of responsibility — 0.75
Computed from structural-signature embeddings · 2026-10-08
Not to Be Confused With¶
Digitization converts an existing analog representation to digital form; datafication renders a practice or process as processable data. Informating is digital work's simultaneous informational by-product. Informatization is systemic growth of information activity. Dataism is the belief in the objectivity or comprehensiveness of data-driven tracking. Surveillance is one possible use or effect, not the full definition. Prediction, commercial reuse, and asymmetric control must be established locally. A trace of platform use is not automatically a representative measure of all behavior outside that platform.[1][2]
References¶
[1] José van Dijck, “Datafication, dataism and dataveillance: Big Data between scientific paradigm and ideology”, Surveillance & Society 12, no. 2 (2014): 197–208, especially printed pp. 198–200 (PDF pp. 3–5). The repository cover has a “Datafiction” metadata typo; the printed article title is used here. Primary scholarly analysis for the social-platform definition, examples, and limits on treating platform data as neutral or representative. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k ↩l ↩m ↩n ↩o ↩p ↩q ↩r ↩s ↩t
[2] Luci Pangrazio, Neil Selwyn, and Bronwyn Cumbo, “Tracking technology: exploring student experiences of school datafication”, Cambridge Journal of Education 53, no. 6 (2023): 847–862, doi:10.1080/0305764X.2023.2215194, especially printed pp. 847, 849–852 (PDF pp. 2, 5–8). Primary qualitative study of three Victorian secondary schools and 62 students; students' reports and speculations are not a full technical audit of the schools' data flows. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k ↩l ↩m ↩n ↩o ↩p ↩q ↩r ↩s