Phlow System Dashboards
Internal dashboards used to keep Phlow pedagogically rigorous, responsive, and fair as the platform scales.
Phlow System Dashboards
Phlow System Dashboards are not visible to students, parents, or schools. They exist to help Phlow Academy learn about itself.
Their purpose is to evaluate learning design assumptions, calibrate cognitive demand, refine support strategies, improve sequencing and pacing, and ensure the platform remains pedagogically sound as it scales.
These dashboards treat Phlow as a learning system rather than a content delivery mechanism, using evidence from real learning behaviour to improve design decisions over time.
Phlow Dashboard 1: Decision Difficulty Calibration (BDV Review)
This dashboard evaluates whether Base Decision Values accurately reflect real-world cognitive demand. It compares expected difficulty against observed learner effort across many contexts, revealing where decision types may be under- or over-calibrated.
By examining patterns rather than isolated cases, the dashboard supports ongoing refinement of BDV assignments. This ensures that cognitive demand is grounded in evidence rather than assumption, keeping challenge aligned with how learners actually experience decisions.
Phlow Dashboard 2: Decision Stability vs Progression Outcomes
This dashboard explores the relationship between stability signals and successful progression. It helps validate whether readiness thresholds are appropriately set by examining what happens after learners move on.
If progression occurs too early, fragility may emerge; if it is delayed unnecessarily, momentum may be lost. By analysing stability alongside subsequent performance, the dashboard supports careful tuning of progression rules so that readiness reflects genuine understanding rather than short-term success.
Phlow Dashboard 3: Error Pattern Distribution
This dashboard aggregates error patterns across the platform to identify common breakdowns in understanding. Rather than treating errors as isolated mistakes, it examines their distribution across decision types, Phlow structures, and stages.
This reveals where conceptual misunderstandings, execution slips, or cognitive overload are most likely to occur. These insights inform redesign of prompts, sequencing, and support — improving learning design at its source rather than remediating downstream symptoms.
Phlow Dashboard 4: Effectiveness of Support Interventions
This dashboard evaluates which support mechanisms genuinely improve learning stability over time. It distinguishes between supports that merely increase immediate correctness and those that lead to lasting understanding.
By comparing learning behaviour before and after different interventions, the dashboard helps identify which scaffolds should be strengthened, refined, or retired. Support design becomes evidence-led, ensuring that assistance promotes independence rather than dependence.
Phlow Dashboard 5: Learning Journey Path Analysis
This dashboard examines how learners move through the platform at a system level. It reveals common pathways, detours, and bottlenecks without exposing individual trajectories.
By analysing these aggregated journeys, Phlow can identify where sequencing supports flow and where it creates friction. This enables continuous refinement of learning pathways so that progression remains coherent, adaptive, and humane at scale.
Phlow Dashboard 6: Cognitive Load Distribution Across Curriculum
This dashboard identifies where cognitive demand clusters across levels, topics, and Phlow types. By aggregating decision demand across the curriculum, it reveals peaks that may contribute to overload and troughs that may under-challenge learners.
These insights support curriculum balance and pacing decisions, ensuring that difficulty rises intentionally rather than accidentally. Cognitive load is treated as a system property, not an individual failing.
Phlow Dashboard 7: Support Dependency Signals
This dashboard examines how learners rely on support over time and whether that reliance fades appropriately as understanding stabilises. Persistent dependence may indicate over-scaffolding, while rapid withdrawal may signal insufficient support.
By tracking support usage alongside stability measures, the dashboard helps refine fade-out rules so that assistance is removed neither too early nor too late. The goal is productive independence, not unsupported struggle.
Phlow Dashboard 8: Learner Profile Evolution
This dashboard tracks how behavioural learner profiles emerge, evolve, and dissolve over time. Profiles are not fixed categories but dynamic patterns that shift as learning develops.
By monitoring profile transitions, Phlow ensures that its understanding of learners remains adaptive rather than reductive. This protects against labelling effects and supports a system that responds to growth rather than locking learners into static identities.
Phlow Dashboard 9: Early Warning Design Signals
This dashboard surfaces early indicators that a Phlow, decision type, or sequence may not be functioning as intended. These signals prompt investigation before issues scale widely, allowing design problems to be addressed proactively.
Rather than blaming learners for difficulty, the dashboard treats unexpected patterns as hypotheses about design quality. Attention is directed to where refinement is most needed, supporting continuous improvement.
Phlow Dashboard 10: System Learning Confidence
This dashboard reflects the system’s confidence in its own analytics and recommendations. It indicates where conclusions are strongly supported by accumulated evidence and where signals remain exploratory.
By making analytic confidence visible internally, Phlow avoids overclaiming and supports responsible iteration. Decisions about progression rules, supports, and design changes are guided not just by patterns, but by how reliable those patterns are.
Positioning Note for the Website
This section should be framed as dashboards used internally to ensure Phlow Academy remains pedagogically rigorous, responsive, and fair as it grows.
They demonstrate that analytics are not merely extracted from learners, but reinvested into better learning design. The platform learns alongside its users, continuously refining how learning is supported rather than freezing assumptions in code.