layhadarries/Cape-Town-Suburb-Energy-Vulnerability-Index
Domain:
environment and energygeospatialRecord type:
dataset
Creator:
lay
Host:
A data pipeline to extract dam level and electricity/ loadshedding data from EskomSePush and Cape Town Open Data API. Output is a Google Data Studio report, providing insight into the current loadshedding schedule and dam levels across South Africa.
# Cape Town Suburb Energy Vulnerability Index
A data engineering pipeline that scores Cape Town suburbs by how exposed they are to load shedding — combining actual outage hours with property value, commercial/zoning density, and public facility density — to answer a genuinely useful question: **which suburbs get hit hardest by load shedding, relative to what's actually at stake there?**
Two suburbs can have identical loadshedding stages and hours, but if one has three clinics and a commercially-zoned business strip and the other is mostly low-density residential, the real-world impact is completely different. This project builds a composite index that captures that difference, not just raw outage hours.
## The question this answers
> For each suburb: how many hours of load shedding does it actually experience, and how much economic/civic activity sits behind that suburb (property value, commercial zoning density, public facilities)? Combine both into a single comparable score.
## Architecture
```
┌─────────────────────┐ ┌────────────────────────────┐
│ EskomSePush API │ │ CT Open Data Portal │
│ (outage hours per │ │ (property valuations, │
│ tracked suburb) │ │ zoning, civic facilities)│
└──────────┬──────────┘ └─────────────┬──────────────┘
│ │
└───────────────┬────────────┘
▼
AWS S3 (raw landing zone)
│
▼
AWS RDS Postgres (warehouse)
dim_suburb + fact tables
│
▼
dbt (build vulnerability
index mart, normalize,
weight, test)
│
▼
Looker Studio (suburb map +
ranked score table)
All of the above (except Looker Studio) is orchestrated
by Apache Airflow, running in Docker.
```
## Tech stack
| Layer | Tool |
|---|---|
| Language | Python 3 |
| Extraction | `requests` (CT Open Data), custom EskomSePush client |
| Raw storage | AWS S3 |
| Warehouse | AWS RDS (Postgres) |
| Transformation | dbt-core (`dbt-postgres`) |
| Orchestration | Apache Airflow (Docker Compose) |
| Visualization | Looker Studio |
| …