Chapter 11

Data Hub & Analytics

Market data, construction indices, economic indicators, ML insights, ETL pipelines, spiders, and forecasting.

31 screenshots

Step-by-step guide

Chapter 11 -- Data Hub & Analytics

Overview

The PROPMETRIK Data Hub is the intelligence engine that powers every analytical feature in the platform. It ingests real estate data from government registries, financial institutions, partner APIs, community contributions, and public web sources, then processes it through ETL pipelines, quality checks, and machine-learning models to produce market intelligence for the Ghanaian property market.

The Data Hub and Analytics modules are split across two areas of the platform:

Analytics (Dashboard > Analytics) -- consumer-facing dashboards for market insights, forecasting, risk, affordability, construction costs, and ML model outputs.
Data Hub Administration (Dashboard > Admin > Data Hub) -- backend administration for data sources, ingestion pipelines, quality management, spiders, and system configuration.

All analytics use PROPMETRIK's terminal-style dark UI with amber accents and monospaced typography.

Analytics main dashboard showing market KPIs, price indices, and supply-demand metrics

Part A: Analytics Dashboards

11.1 Market Analytics (Main)

Navigate to Dashboard > Analytics to access the main market intelligence dashboard.

KPI Summary Bar

Six metric panels across the top:

Panel	Shows
AVG PRICE	Average property price across all regions (GH₵)
YoY GROWTH	Year-over-year price change (green positive, red negative)
MONTHLY VOLUME	Transaction count for the current period
ACTIVE LISTINGS	Properties currently on the market
NEW LISTINGS	Properties listed this period
MARKET STATUS	Temperature reading (Hot, Warm, Balanced, Cool, Cold) with months of supply

Area Price Index

A table showing price indices by region and property type:

Region name
Property type (Residential, Commercial, Land, etc.)
Average price (GH₵)
Median price
Year-over-year change
Transaction volume

Recent Transactions

A sidebar panel listing the latest property transactions with:

Location
Property type and bedrooms
Final price in GH₵
Price per square metre
Transaction date and type (sale/rental)

Supply & Demand by Region

A table breaking down each region's market dynamics:

Total supply
Properties sold in the last 30 days
Inventory months (how long current supply would last at current sales pace)
Average days on market
Market temperature (colour-coded: red for hot, amber for warm, green for balanced, blue for cool)

Price Distribution

A horizontal bar chart showing the percentage of transactions in each price bracket (e.g. "GH₵ 100K--200K: 23%").

Price Index Trend (12 Months)

A sparkline bar chart showing monthly median prices and month-over-month change percentages for the trailing 12 months.

Market Activity by Region

A comprehensive table with:

Transaction count and total value
Average price
Active and new listings
Average days on market
Average discount from asking price

Tip: Click the INVESTMENT FINDER button in the header to jump to the Investment Analysis module for buy-side recommendations.

11.2 Affordability Analytics

Navigate to Dashboard > Analytics > Affordability.

Affordability analytics showing GHAI composite scores, regional breakdowns, and trend charts

The Ghana Housing Affordability Index (GHAI) is a PROPMETRIK-exclusive composite index that measures how affordable housing is across regions.

GHAI Components

Component	Abbreviation	Description
Mortgage Housing Affordability Index	MHAI	Affordability via mortgage financing
Cash Housing Affordability Index	CHAI	Affordability for outright cash purchases
Rental Housing Affordability Index	RHAI	Rental affordability relative to income
Cost Affordability Index	CAI	Construction cost vs. income
Location Affordability Index	LAI	Transport and infrastructure costs
Market Accessibility Score	MAS	How easy it is to enter the market

Regional Dashboard

For each region the dashboard shows:

GHAI composite score with category (Affordable, Moderate, Stretched, Unaffordable)
Individual component scores
Median property price and household income
Current mortgage rate
Median monthly rent
Trend direction and month-over-month / year-over-year changes

Regional Heatmap

A colour-coded map of Ghana showing affordability by region. Darker colours indicate less affordable areas.

Alerts

The system generates alerts when:

A region's GHAI changes category (e.g. Moderate to Stretched)
Mortgage rates change significantly
A region's affordability deteriorates rapidly

11.3 Construction Cost Analytics

Navigate to Dashboard > Analytics > Construction.

Construction cost analytics showing CCI national index, material prices, and regional multipliers

Construction Cost Index (CCI)

The national CCI is a weighted composite of three components:

Materials -- cost of building materials (cement, steel, timber, etc.)
Labour -- skilled and unskilled construction labour rates
Overhead -- site management, insurance, permits

Each component shows:

Current index value
Component weight in the composite
Month-over-month change
Year-over-year change

Regional Cost Multipliers

A table showing how costs vary by region:

Regional multiplier (1.0 = national average)
Distance from port (affects transport costs)
Transport factor
Infrastructure score
Cost component breakdown (materials, labour, transport)
6-month and 12-month trends

Material Prices

A detailed table of current material prices:

Category and sub-category
Unit price in GH₵ with unit of measure
Local vs. import percentage
Price volatility rating (Low, Medium, High)
Month-over-month and year-over-year changes

Tip: Use the CCI trend chart to time large material purchases -- buying during seasonal dips can save 5--15% on material costs.

11.4 Forecasting

Navigate to Dashboard > Analytics > Forecasting.

Forecasting dashboard showing CCI and GHAI predictions with confidence intervals

CCI Forecasting

The Construction Cost Index forecaster projects future cost indices:

Forecast Horizon -- 3, 6, or 12 months ahead.
Predicted Index -- the projected CCI value.
Confidence Interval -- upper and lower bounds.
Model Info -- method used, data points, R-squared, trend direction, monthly change rate.
Component Forecasts -- separate projections for materials, labour, and overhead.

GHAI Forecasting

The affordability forecaster projects how affordability will change:

Current GHAI and category
Predicted GHAI at each future period
Whether the category is expected to change (e.g. Moderate to Stretched)
Contributing factors driving the forecast

Price Forecasting

Property price predictions by region:

Current median price
3/6/12-month price projections
Confidence intervals
Key drivers (supply changes, economic indicators, construction costs)

11.5 Geographic Analytics

Navigate to Dashboard > Analytics > Geographic.

Geographic analytics showing regional comparison maps and neighbourhood-level metrics

Interactive regional comparison charts
Neighbourhood-level metrics (average price, transaction volume, days on market)
Urban vs. peri-urban analysis
Infrastructure proximity scoring
School and amenity accessibility indices

11.6 Risk Analytics

Navigate to Dashboard > Analytics > Risk.

Flood Risk

Flood risk scores by neighbourhood (0--100 scale)
Risk level classification (Low, Moderate, High, Very High)
Nearby flood incidents with distance
Risk factors list
Zone type and severity
Recommendations

Recent Flood Incidents

A timeline of flood events with:

Source and date
Description and severity
Neighbourhood and city
Latitude/longitude coordinates

Land Litigation Risk

Active cases by region
Case reference and dispute type
Status (pending, resolved, appealed)
Risk score
Filing date

Environmental Hazards

Proximity to industrial zones
Soil contamination risk
Noise pollution zones
Air quality indices

Tip: Always check the Risk Analytics for a property's neighbourhood before finalising a valuation or deal -- flood risk alone can impact value by 10--25%.

11.7 Short-Stay Analytics

Navigate to Dashboard > Analytics > Short-Stay.

Short-stay analytics showing Airbnb-style rental performance metrics and occupancy rates

Average daily rates by neighbourhood
Occupancy rates
Revenue per available night
Seasonal demand patterns
Competition analysis (number of listings by area)
Guest review score averages

11.8 Investment Analysis

Navigate to Dashboard > Analytics > Market > Investments.

Investment analysis showing yield calculations, cap rates, and ROI projections

Gross and net yield calculations by region
Capitalisation rates
Cash-on-cash return projections
Break-even analysis
Comparative investment scoring
Buy vs. rent decision framework

11.9 Rental Market Analysis

Navigate to Dashboard > Analytics > Market > Rentals.

Rental market analysis showing rental trends, vacancy rates, and yield maps

Median rent by region and property type
Rental yield trends
Vacancy rates
Rent growth year-over-year
Landlord return analysis
Rental demand indicators

11.10 ML Insights Dashboard

Navigate to Dashboard > Analytics > ML.

ML insights dashboard showing model performance, confidence distribution, and drift detection

The ML (Machine Learning) dashboard provides transparency into the predictive models powering PROPMETRIK's analytics.

Service Status

ML service status (online/offline)
Model version in production
Total predictions in the last 30 days
Average confidence score
Active drift alerts

Model Performance Metrics

Metric	Description
MAE	Mean Absolute Error -- average prediction error in GH₵
RMSE	Root Mean Squared Error -- penalises large errors more
MAPE	Mean Absolute Percentage Error
R-squared	Proportion of variance explained (closer to 1.0 is better)
Median Error	Middle error value
P90 Error	90th percentile error (worst 10% of predictions)
Within 10%	Percentage of predictions within 10% of actual values
Within 20%	Percentage of predictions within 20% of actual values

Ensemble Analytics

PROPMETRIK uses an ensemble of multiple models. The dashboard shows:

Individual model weights
Model agreement scores
Ensemble vs. individual model performance

11.11 ML Feature Importance

Navigate to Dashboard > Analytics > ML > Features.

ML feature importance showing ranked feature contributions and SHAP values

Ranked list of features (inputs) used by the valuation model
Feature importance scores
SHAP (Shapley) values showing how each feature pushes predictions up or down
Feature distribution histograms
Correlation matrix between key features

11.12 ML Forecasting

Navigate to Dashboard > Analytics > ML > Forecasting.

ML forecasting showing time-series predictions with confidence bands

Property value time-series forecasts
Confidence bands (narrow = high confidence, wide = uncertain)
Historical accuracy tracking
Model comparison (which model performed best by region/property type)

11.13 ML Model Monitoring

Navigate to Dashboard > Analytics > ML > Monitoring.

Drift Detection

Monitors whether the model's input data distribution has shifted:

Concept Drift -- the relationship between inputs and outputs has changed.
Data Drift -- the distribution of incoming data has shifted.
Prediction Drift -- model outputs are systematically different from expectations.

Each drift alert includes:

Detection date
Drift type and severity
Affected metrics
Recommendation (retrain, monitor, or no action)
Whether retraining is required

Performance Trends

Track how model accuracy changes over time:

Trend direction (improving, degrading, stable)
Change rate
Historical performance charts

Confidence Distribution

Histogram of prediction confidence scores
High/medium/low confidence breakdown
Mean and median confidence over time

Part B: Data Hub Administration

11.14 Data Hub Overview

Navigate to Dashboard > Admin > Data Hub for the administration dashboard.

Data Hub main dashboard showing source statistics, job status, and real-time data feed

The Data Hub overview displays:

Source Statistics -- total data sources by tier, active vs. paused.
ETL Job Status -- running, queued, completed, and failed jobs.
Pending Contributions -- community-submitted data awaiting review.
Economic Snapshot -- latest macro-economic indicators.
Data Quality Widget -- overall quality scores and alerts.
Real-Time Data Feed -- live stream of incoming data events.
System Health -- server status, queue depths, processing rates.

Data Source Tiers

PROPMETRIK organises data sources into five tiers based on reliability:

Tier	Name	Examples	Colour
Tier 1	Government	Lands Commission, Ghana Statistical Service	Blue
Tier 2	Financial	Banks, mortgage providers, insurance companies	Green
Tier 3	Partners	Estate agents, developers, professional bodies	Purple
Tier 4	Contributions	User-submitted transactions and market data	Amber
Tier 5	Public Web	Property listing websites, classifieds	Orange

11.15 Data Sources

Navigate to Admin > Data Hub > Sources.

Data sources page showing source list with tier badges, sync status, and health indicators

Viewing Sources

The source list shows:

Source name and description
Tier badge (colour-coded)
Status (Active, Paused, Inactive)
Last sync timestamp
Record count
Health indicator (green checkmark, yellow warning, red error)

Filtering

Search -- find sources by name or description.
Tier filter -- show only sources of a specific tier.
Status filter -- filter by Active, Paused, or Inactive.

Source Actions

Play/Pause -- toggle a source between active and paused states.
Refresh -- trigger an immediate data pull from the source.
View Details -- inspect source configuration, API endpoints, and error logs.

11.16 Ingestion

Navigate to Admin > Data Hub > Ingestion.

Ingestion pipeline view showing data flow stages and processing metrics

The ingestion pipeline processes raw data through multiple stages:

Extract -- pull data from the source (API call, file download, web scrape).
Transform -- normalise, deduplicate, geocode, and enrich records.
Validate -- run quality checks (completeness, consistency, accuracy).
Load -- write validated records to the data warehouse.

The ingestion page shows:

Pipeline stages with processing counts
Current throughput (records per minute)
Error rates per stage
Queue depth
Processing latency

11.17 Data Catalog

Navigate to Admin > Data Hub > Catalog.

Data catalog showing dataset registry with schemas, freshness, and usage statistics

The data catalog is a searchable registry of all datasets in the platform:

Dataset name and description
Schema (columns, data types, constraints)
Source tier and origin
Freshness (when last updated)
Record count
Usage statistics (which analytics modules reference this dataset)
Data lineage links

Searching the Catalog

Use the search bar to find datasets by name, column name, or description. Filter by tier or data domain (property, economic, demographic, geographic).

11.18 Data Quality

Navigate to Admin > Data Hub > Quality.

Quality Scores

Each dataset receives a composite quality score based on:

Dimension	Description
Completeness	Percentage of non-null values in required fields
Consistency	Cross-field validation (e.g. price > 0 when status = "sold")
Accuracy	Spot-check accuracy against verified sources
Timeliness	How recent the data is relative to expected refresh cadence
Uniqueness	Percentage of non-duplicate records

Validation Rules

Define custom validation rules for each dataset:

Click + Add Rule.
Select the dataset and field.
Choose the rule type (not null, range, regex, cross-field, custom SQL).
Set the severity (Error, Warning, Info).
Click Save.

Issue Tracking

Failed validation rules generate issues that appear in the issues table with:

Dataset and field affected
Rule that failed
Severity
Record count affected
First and last occurrence
Resolution status

11.19 Data Lineage

Navigate to Admin > Data Hub > Lineage.

Data lineage diagram showing data flow from sources through transformations to analytics outputs

The lineage view provides a visual graph showing how data flows through the system:

Source nodes (Tier 1--5)
ETL transformation steps
Intermediate datasets
Final analytical outputs

Click any node to see:

Input sources
Transformation logic applied
Output destinations
Data freshness at each stage

Tip: Use lineage to debug data issues -- trace a suspicious metric back to its source to identify where the problem originated.

11.20 Data Analytics

Navigate to Admin > Data Hub > Analytics.

Data analytics showing ingestion trends, source performance, and data growth charts

Internal analytics about the data platform itself:

Ingestion volume over time (records per day/week/month)
Source reliability rankings
Data growth rates by tier
Processing time distributions
Error rate trends

11.21 Construction Data

Navigate to Admin > Data Hub > Construction.

Construction data management showing material prices, labour rates, and supplier data

Manage the underlying construction cost data that feeds the CCI:

Material price records with supplier, region, and date
Labour rate entries by skill level and region
Bulk upload interface for periodic price surveys
Historical price charts per material
Supplier directory

11.22 Economic Data

Navigate to Admin > Data Hub > Economic.

Economic data dashboard showing GDP, inflation, exchange rates, and interest rates

Manage macro-economic indicators used in analytics:

GDP growth rates
Inflation (CPI, PPI)
Bank of Ghana policy rate
Commercial mortgage rates
Exchange rates (GH₵/USD, GH₵/EUR, GH₵/GBP)
Unemployment statistics
Building permits issued

Data is sourced from:

Ghana Statistical Service
Bank of Ghana
Ministry of Finance
World Bank
IMF

11.23 Insights

Navigate to Admin > Data Hub > Insights.

Insights page showing auto-generated market intelligence reports and anomaly detection

Auto-generated intelligence reports:

Market trend summaries
Anomaly detection alerts (unusual price movements, transaction spikes)
Regional comparison reports
Seasonal pattern analysis
Investment opportunity signals

11.24 Performance

Navigate to Admin > Data Hub > Performance.

Performance monitoring showing API response times, query latency, and system throughput

System performance metrics:

API response times (p50, p95, p99)
Database query latency
Cache hit rates
Concurrent users
System throughput (requests per second)
Resource utilisation (CPU, memory, disk)

11.25 ETL Jobs

Navigate to Admin > Data Hub > Jobs.

ETL jobs list showing job status, progress bars, and execution history

Job List

All ETL jobs with:

Job name and type (Extract, Transform, Load, Full Pipeline)
Status badge (Running, Queued, Completed, Failed, Cancelled)
Progress bar (for running jobs)
Start time and duration
Records processed
Error count

Filtering

Status filter -- show only running, completed, or failed jobs.
Type filter -- filter by job type.

Job Actions

Cancel -- stop a running job with an optional reason.
View Logs -- inspect the terminal-style log output.
Retry -- re-run a failed job.

Job Statistics

Summary cards show:

Total jobs run
Currently running
Success rate
Average duration
Jobs in queue

11.26 Queues

Navigate to Admin > Data Hub > Queues.

Queue management showing processing queues with depth, throughput, and worker status

Monitor and manage background processing queues:

Queue name and purpose
Current depth (messages waiting)
Processing rate (messages per minute)
Worker count and status
Dead letter queue size (messages that failed all retry attempts)
Oldest message age

Queue actions:

Pause/Resume -- temporarily halt processing.
Purge -- clear all messages from the queue (confirmation required).
Retry Dead Letters -- re-queue failed messages for another attempt.

11.27 Pull Integrations

Navigate to Admin > Data Hub > Pull Integrations.

Pull integrations showing configured API connections with sync schedules and status

Manage scheduled data pulls from external APIs:

Integration name and target API
Sync schedule (cron expression)
Last sync time and status
Records fetched
Authentication status
Configuration (API keys, endpoints, pagination settings)

To add a new integration:

Click + New Integration.
Enter the API endpoint URL and authentication credentials.
Configure the data mapping (source fields to PROPMETRIK fields).
Set the sync schedule.
Test the connection.
Click Save & Activate.

11.28 Web Spiders

Navigate to Admin > Data Hub > Spiders.

Web spiders crawl property listing websites to collect market data. Each spider targets a specific website or data source.

Spider List

Each spider shows:

Spider name and target website
Status badge (Idle, Running, Failed, Paused)
Pages crawled and records extracted
Last run time and duration
Success rate
Schedule

Spider Actions

Play -- start a spider crawl immediately.
Pause -- suspend a running crawl.
Stop -- terminate a running crawl.
Refresh -- update the status display.

Spider Statistics

Total pages crawled
Records extracted
Error rates
Average pages per minute
Data freshness

Tip: Schedule spiders to run during off-peak hours (midnight to 6 AM WAT) to minimise impact on target websites and reduce blocking risk.

11.29 Contributions

Navigate to Admin > Data Hub > Contributions.

Community members and partner agents can submit property transaction data. The contributions page manages the review workflow:

Review Queue

Pending contributions with:

Submitter name and organisation
Property details (location, type, price)
Transaction date
Supporting evidence (photos, documents)
Confidence score (auto-assessed)

Review Actions

Click a contribution to view full details.
Cross-reference against existing records.
Choose an action:
- Approve -- accept the data into the platform.
- Reject -- decline with a reason.
- Request More Info -- ask the submitter for additional details.
- Flag for Review -- escalate to a senior data analyst.

Contributor Leaderboard

Track top contributors:

Number of submissions
Approval rate
Data quality score
Contributor tier (Bronze, Silver, Gold, Platinum)

11.30 Valuation Configuration

Navigate to Admin > Data Hub > Valuation Config.

Valuation configuration showing model parameters, comparable selection rules, and adjustment factors

Configure the parameters that drive automated valuation models:

Comparable Selection Rules -- radius, property type matching, age limits, size variance.
Adjustment Factors -- per-feature adjustments (e.g. +5% for corner plot, -3% per year of age difference).
Model Weights -- relative importance of comparable sales, cost approach, and income approach.
Confidence Thresholds -- minimum confidence score to auto-approve a valuation.
Regional Overrides -- custom parameters for specific regions.

11.31 Data Hub Settings

Navigate to Admin > Data Hub > Settings.

Data Hub settings showing system configuration, API keys, and notification preferences

System-wide Data Hub configuration:

API Rate Limits -- maximum requests per minute for each tier.
Retention Policies -- how long to keep raw data, processed data, and logs.
Notification Settings -- email/Slack alerts for job failures, quality issues, and drift detection.
Access Control -- which roles can view, edit, or administer data hub features.
Backup Configuration -- automated backup schedules and retention.
Environment Variables -- API keys and secrets for external integrations (masked in the UI).

Summary

The Data Hub & Analytics module is the backbone of PROPMETRIK's market intelligence capabilities. The analytics dashboards transform raw real estate data into actionable insights -- from affordability indices and construction cost trends to ML-powered valuations and risk assessments. The administration tools give data teams full control over the ingestion pipeline, quality management, and system health. Together, they provide a comprehensive data platform purpose-built for the Ghanaian real estate market.