Open-Sourcing the Causal Climate Dataset Scripts for Kenya’s ASALs

A GreenScope Analytics Platform Test

Why We Open-Sourced This

At GreenScope Analytics, we are building a platform for causal climate intelligence, not just climate prediction.

As part of stress-testing the platform, we made a deliberate decision to open-source the code used to construct the core causal dataset for Kenya’s Arid and Semi-Arid Lands (ASALs).

This is not the entire platform.
It is the foundation layer — what we internally refer to as Module 0.

Why open-source this layer?

Because if the data itself is not:

transparent,

reproducible,

and causally coherent,

then no amount of modeling on top of it can be trusted.

---

What Exactly Is Open-Sourced?

The open-sourced repository contains the code used to build a causal-ready climate dataset, starting from raw global and local observations and ending with a unified ASAL data cube.

Specifically, the repository includes:

Data ingestion and validation scripts

Spatial and temporal harmonization logic

Memory-efficient handling of large climate files

Explicit causal role assignment across variables

This code produces a dataset designed for causal analysis, not just correlation-based modeling.

👉 The repository focuses on how the dataset is constructed, not on proprietary modeling, analytics, or downstream decision layers.

---

The Datasets Involved

The open-sourced pipeline integrates:

Global Climate Drivers (Exogenous)

ENSO (El Niño–Southern Oscillation) — Niño 3.4 Index (NOAA)

Indian Ocean Dipole (IOD) — Dipole Mode Index (NOAA)

These are treated as external forcing mechanisms, capable of triggering large-scale climate regime shifts.

Local Climate & Ecological Systems

CHIRPS – Precipitation (water input)

ERA5-Land – Potential Evaporation & Soil Moisture (atmospheric demand and hydrological memory)

MODIS NDVI – Vegetation response

Each variable is assigned a causal role (driver, stressor, mediator, or response) at the dataset level — before any modeling begins.

---

Why This Is Not “Just Research Code”

This work is part of a platform test, not a standalone academic exercise.

The objective was to answer a very practical question:

Can GreenScope’s infrastructure reliably turn heterogeneous climate data into a causally structured dataset at scale?

By open-sourcing this step, we are:

Pressure-testing our data engineering assumptions

Inviting scrutiny of our causal design choices

Demonstrating how we handle real-world climate data constraints

Providing a reference implementation for others working on climate causality

In short: this is production-grade thinking, even though the platform itself remains private.

---

What We Did Not Open-Source (Intentionally)

To be clear, this repository does not include:

Proprietary causal discovery algorithms

Intervention or counterfactual engines

Decision-support logic

Platform orchestration layers

Those remain part of the GreenScope Analytics platform.

The goal here is trust at the data foundation, not exposure of the full stack.

---

Why This Matters

Climate risk in regions like Kenya’s ASALs is driven by:

non-stationary relationships,

interacting global and local systems,

and delayed, cascading effects.

If we want climate intelligence systems that are:

explainable,

trustworthy,

and actionable,

then causal datasets must be treated as first-class infrastructure — not as an afterthought.

This open-source release is our contribution toward that direction.

---

Where to Find the Code

The repository is publicly available on GitHub and documents:

How raw climate datasets are ingested

How they are aligned spatially and temporally

How causal structure is preserved at the data level

How large files are handled safely and reproducibly

Github Repo: (https://github.com/LeahN67/greenscope-asal-casuality-dataset)

📌 This repository represents the dataset construction layer used to test the GreenScope Analytics platform for ASAL climate intelligence.

---

Closing

Open-sourcing this work is not about giving everything away.
It is about earning trust where it matters most — at the data layer.

We’re deep in the trenches building this platform, and this release reflects how seriously we take that responsibility.