How a Climate Technology Organisation Tracked Carbon Removal in Real Time

A climate technology organisation runs small-scale, decentralised biochar production in underserved communities. As deployments grew, it had no way to see production or net carbon removal as it happened, and manual tracking could not keep pace. OnPoint Insights built a cloud-native platform on Azure Data Lake Storage and Databricks that collects sensor data continuously and shows carbon removal site by site.

Manual reporting replaced

Production and net carbon removal readings now arrive continuously from IoT sensors through the Carbon Pilot API.

Site by site, near real time

Carbon removal, production rates and threshold breaches are visible per deployment as they happen, not at the end of a reporting period.

Breaches caught at the site

Threshold breaches surface in the dashboard, so a local leader can act on their own deployment instead of waiting for a central review.

The problem

Impact you cannot measure is impact you cannot prove

Carbon removal only counts when it can be evidenced. For a decentralised operation that means knowing what each site produced and removed, continuously, rather than reconstructing it at the end of a reporting period.

Our client produces biochar through small, decentralised units in underserved communities, with environmental justice as the point of the work. As deployments multiplied, the data behind the impact did not keep up.

What was getting in the way

No central system monitored biochar production or net carbon removal in real time.
Data collected across deployments was inconsistent, which made analysis difficult.
Manual tracking could not keep pace as the number of sites grew.
Emissions data was becoming harder to manage at scale, just as scale became the goal.

Our approach

Collect it automatically, then make it provable

What we found

The organisation was not short of commitment or of sensors. The gap sat between the sensor and the decision: readings arrived in different shapes from each deployment, and any manual step in the middle would break the moment another site came online.

What we did

Data that collects itself. The Carbon Pilot API ingests biochar production and net carbon removal readings from IoT sensors continuously, replacing manual reporting from each site.

One place for everything, structured or not. Azure Data Lake Storage holds structured and unstructured emissions data from every location, so analysis starts from a single source rather than a reconciliation.

Analytics that look forward, not only back. Data is cleaned, processed and modelled in Databricks to support forecasting and sustainability planning, not just reporting on what already happened.

Dashboards for the people on the ground. Interactive Databricks dashboards show site-level carbon removal, threshold breaches and sustainability KPIs, for central decision-makers and local leaders alike.

The choice we made

We built on a data lake rather than a fixed database schema. Readings from a new deployment can land in whatever shape that site’s sensors produce, which is what lets the organisation add rural sites without redesigning the platform each time.

BUILT ON

Azure Data Lake Storage
Databricks

The results

What changed for the organisation

Sustainability reporting stopped being a periodic exercise and became a live view of what each site is actually removing.

Carbon removal is visible as it happens

Emissions, carbon removal and production rates can be monitored live across locations, instead of being compiled after the fact from site records.

Problems get caught at the site

Threshold breaches surface in the dashboard, so a local leader can act on their own site rather than wait for a central review to notice.

Reporting that community leaders and researchers can trust

Transparent, visual performance metrics mean the numbers behind the impact claims can be seen by the people the work is for, which matters for an organisation built on environmental justice.

Delivered by a Microsoft Solutions Partner for Data & AI

Let's Connect