Use Cases

The use case model defines all interactions between the four actors (End User, Data Contributor, Administrator, System) and the platform. It is divided into two sets: Observation Prediction (the core analytical domain) and User Management (access control and alerting).

---

Observation Prediction use case diagram

This diagram maps the Admin and User actors to the full set of spatial data and forecast operations.

Actor responsibilities

Actor	Use cases
Admin	Create Zone, Read Zone, Update Zone (includes Read Zone)
User	Create/Read Malaria_Observation, Create/Read NCD_Observation, Create/Read Ground_Observation
System/ML	Create Malaria_Risk_Forecast, Create NCD_Risk_Forecast
User/System	Read Malaria_Risk_Forecast, Read NCD_Risk_Forecast

---

User Management use case diagram

This diagram maps the Admin and User actors to role, user profile, and alert management operations.

Actor responsibilities

Actor	Use cases
Admin	Create Role, Read Role, Update Role (includes Read Role); Create User
User	Read User (includes Read Role), Update User (includes Read User)
User	Create Alert (includes Read User), Read Alert (includes Read User), Update Alert Status (includes Read User + Read Alert)
System	Create Alert automatically on high-risk forecast

---

Use case descriptions

Create Malaria_Observation

Attribute	Detail
Trigger	New 8-day satellite imagery batch available from GEE
Actors	System Integration Service, External API (GEE)
Pre-conditions	ObservationCount > Minimum Threshold; ML model weights loaded
Flow	1. GEE API generates latest imagery → 2. System pulls data for zone coordinates → 3. GEE returns JSON (temperature, soil moisture, water index) → 4. System maps to zone_id → 5. DB validates zone → 6. System inserts Malaria_Observation row
Post-condition	New record in MALARIA_OBSERVATION: [env_record_id, zone_id, date_recorded, lst_surface_temp_c, soil_moisture_pct]

---

Create Malaria_Risk_Forecast

Attribute	Detail
Trigger	Scheduled integration service fires (every 8 days on new satellite pass)
Actors	ML Pipeline
Pre-conditions	Sufficient historical training data; ML model weights active
Flow	1. ML Pipeline initiates prediction sequence → 2. Queries lst_surface_temp_c, soil_moisture_pct (READ Malaria_Observation) → 3. Queries aggregated_case_count (READ Ground_Observation) → 4. Applies environmental_lag_days, runs XGBoost → 5. Outputs outbreak_probability_score per zone → 6. DB validates zone_id → 7. INSERT into MALARIA_RISK_FORECAST → IF score > 0.80 → trigger Create User_Risk_Alert
Post-condition	New record in MALARIA_RISK_FORECAST: [forecast_id, zone_id, target_prediction_date, outbreak_probability_score]

---

Create User_Risk_Alert

Attribute	Detail
Trigger	ML Pipeline saves forecast with outbreak_probability_score > 0.80
Actors	System (automated)
Pre-conditions	High-risk forecast exists
Flow	1. System intercepts high-risk trigger → 2. Queries all users assigned to affected zone_id → 3. Retrieves user_id and email_address from USER table → 4. Formats alert payload (status = 'Active') → 5. Dispatches email → 6. INSERT into USER_RISK_ALERT
Post-condition	New records in USER_RISK_ALERT linking user_id to forecast_id with alert_status = 'Active'

---

Update Alert Status

Attribute	Detail
Trigger	User clicks "Acknowledge" or "Resolve" on an active alert in the dashboard
Actors	User
Pre-conditions	Alert exists in DB with alert_status = 'Sent' or 'Active'; user authenticated
Flow	1. User logs in, navigates to Active Alerts panel → 2. System fetches alerts for user's zones → 3. User clicks alert → 4. System fetches corresponding forecast details → 5. User reviews outbreak_probability_score and map → 6. User clicks 'Acknowledge Alert' → 7. System updates alert_status = 'Acknowledged'
Post-condition	Alert record updated; removed from Active view; moved to Historical/Resolved

---

Read Malaria_Risk_Forecast

Attribute	Detail
Trigger	User navigates to Risk Map dashboard view
Actors	User
Pre-conditions	Forecasts exist for the current date range
Flow	1. User clicks Risk Map tab → 2. System requests polygon boundaries (ZONE table) and latest forecasts (MALARIA_RISK_FORECAST) → 3. System merges spatial data with forecast scores, applies colour coding (red > 0.80) → 4. User views rendered interactive map
Post-condition	Read-only — no DB state changes

---

Download use case document

Download Use Case Set (.docx) →