Kings Cross Seasonal Demand Intelligence

Urban demand volatility detection & seasonal pattern learning Kings Cross / Coal Drops Yard · London

---

Status: Seasonal Data Collection (Paused Development)

This project is currently in data-collection mode.

From 22 December 2025 to mid-January 2026, no new features, UI changes, or anomaly logic are being added.

Purpose of the pause

Capture clean seasonal demand volatility signals under a frozen model, allowing post-period analysis without confounding changes.

Dataset lock

• Model & anomaly logic frozen: 22 Dec 2025 → 7 Jan 2026
• All insights derive from an unchanged seasonal model
• Any findings are attributable to real-world conditions, not code iteration

This pause is intentional and foundational for downstream insight quality.

---

Overview

Kings Cross Seasonal Demand Intelligence is an experimental urban analytics system designed to:

• Detect demand volatility in high-footfall districts
• Explain why deviations occur using observable public signals
• Learn seasonal demand behaviour without requiring private data

The system focuses on district-level dynamics, not individual venue performance.

It is built to support:

• Property & asset managers
• Urban operators
• Place-making teams
• PropTech & analytics partners
• Consultants analysing seasonal or event-driven demand risk

---

What This Is (and Is Not)

This is:

• A signal-based demand intelligence engine
• Explainable, taxonomy-driven anomaly detection
• Infrastructure-aware (transport, weather, events)
• Designed for replicability across districts and cities

This is not:

• A POS system
• A sales forecast tool
• A marketing optimisation platform
• A restaurant operations dashboard

The system deliberately avoids private or sensitive data sources.

---

Core Signals

The pipeline ingests public, explainable signals only:

• Weather (temperature, conditions)
• Transport status (TfL disruptions & pressure)
• Local events (Eventbrite)
• Venue density & proximity (Google Places)
• Time-of-day & seasonal context

These are fused into a single district busyness signal and compared against seasonal baselines.

---

Anomaly Taxonomy (v1)

All deviations are classified using a fixed taxonomy:

Demand anomalies

• unexpected_peak
• suppressed_demand
• prolonged_peak
• volatile_demand

Timing anomalies

• shifted_peak
• missing_peak

Signal mismatch anomalies

• transport_demand_mismatch
• weather_demand_mismatch
• event_demand_mismatch

Each anomaly includes:

• Severity (low, medium, high)
• Confidence (signal agreement, not certainty)
• Persistence (transient, emerging, established)
• Human-readable explanation
• Contributing drivers

---

Seasonal Learning Phase (Dec 2025 – Jan 2026)

This period was chosen intentionally due to:

• Christmas trading volatility
• Pre-NYE build-up
• NYE spike
• Post-holiday normalization

During this phase the system collected:

• Continuous hourly observations
• Anomaly persistence patterns
• Driver co-occurrence statistics
• Confidence stability metrics

No tuning or optimisation occurred during collection.

---

Outputs

The system produces:

• kingscross_dashboard.json → Current state, context, venues, cluster pressure

• forecast.json → Short-term baseline demand projection

• history/kingscross_history.json → Long-running demand signal history

• anomalies.json → Fully classified seasonal deviations

• observations.json → Raw signal truth-log for learning & audit

• seasonal_insights_2025.json → Aggregated post-season analysis (counts, patterns, interpretations)

---

Explainability First

Every insight is:

• Traceable to observable signals
• Logged with drivers
• Interpretable by non-technical stakeholders

Confidence reflects signal agreement, not ground truth certainty.

This makes the system suitable for:

• Risk discussions
• Strategic planning
• Post-event analysis
• Investor or board-level reporting

---

Why Kings Cross?

Kings Cross was selected as a validation district because it combines:

• Transport complexity
• Mixed-use footfall
• Event-driven volatility
• Strong seasonal effects

The architecture is not Kings Cross–specific and is designed for reuse.

---

Roadmap (Locked During Pause)

Completed

• Seasonal anomaly engine (v1)
• Persistence tracking
• Explainable drivers
• Frozen seasonal dataset

Next (post-pause)

• Seasonal insight synthesis (Jan report)
• District-to-district replication
• Decision signal abstraction
• Product narrative & buyer positioning

No roadmap items will be executed until the seasonal dataset is fully analysed.

---

Data Ethics & Scope

• No personal data
• No payment data
• No device tracking
• No private venue data

All signals are public, aggregated, and explainable.

---

Project Intent

This project explores whether urban demand volatility itself is a valuable signal — independent of sales, marketing, or venue-level optimisation.

The goal is to determine whether explainable seasonal demand intelligence can support better decision-making at the district and asset level.

---