The challenge

Validating the effectiveness of R-leaf requires field measurements using specialist gas analysis equipment, followed by appropriate validated calculations to determine greenhouse gas removal.
This process presented several challenges:

• Manual, spreadsheet-based workflows that were time-intensive
• Large volumes of raw data requiring significant processing
• Fragmented data collection across field and laboratory environments
• Limited ability to scale analysis or generate predictive insights

To support the scale-up potential of the technology, Crop Intellect sought a more efficient approach that would reduce manual effort and enable future predictive modelling.

The approach

Through the BridgeAI programme, activity was delivered through the National Innovation Centre for Data (NICD) via the Hartree Centre North East Hub, part of the wider Hartree Centre SME Hubs, which provide regional advanced digital technology support to UK industry. NICD worked with Crop Intellect to deliver a project focused on both technical development and capability building.

At the outset, NICD assessed the available data and identified that a single season of field data would not be sufficient to build a robust predictive model. As a result, the project prioritised establishing strong data foundations and enabling the business to scale its analytical capabilities over time.

NICD collaborated closely with the Crop Intellect team through a series of structured, hands-on sessions.

These sessions introduced:

• Python-based data processing and automation
• Development environments using VS Code
• Version control and collaboration using Git and GitHub
• Exploratory data analysis and visualisation using pandas and matplotlib
• Consideration of modelling approaches for future development

A key milestone in the project was the development of a reproducible data pipeline, enabling the transition from manual processing to automated workflows.

Knowledge transfer was embedded throughout, ensuring that Crop Intellect could apply these approaches independently beyond the project.

The solution

NICD and Crop Intellect developed a reproducible data processing pipeline to transform raw LI-COR field measurements into structured, analysis-ready datasets.

The solution included:

• Automated data cleaning and structuring
• Outlier detection and filtering
• Streamlined calculation of gas concentration and flux values
• Preparation of datasets for further analysis and modelling

This replaced manual spreadsheet-based processes with a consistent and scalable workflow, improving both efficiency and reliability.

In parallel, the project explored the use of environmental variables, including weather data, to inform predictive modelling. While further data collection is required, this work established a clear direction for future development.

Outcomes and impact

The project delivered measurable benefits for Crop Intellect, supporting both immediate operational improvements and longer-term innovation. Automating data processing has reduced manual effort and increased the speed and consistency of analysis.

“The biggest impact would be the efficiency… it got easier to process the data.” Yusuf Khambhati, Research Scientist, Crop Intellect

Alongside this, the business has adopted more structured approaches to data collection, organisation, and processing, addressing previous challenges with fragmented systems and improving overall data usability.

The project also contributed to increased internal capability, with the development of in-house expertise in data processing and analysis, enabling Crop Intellect to continue building on the work independently and to apply these approaches across future projects.

Importantly, the collaboration has defined a clear pathway towards predictive modelling. By identifying the data requirements needed to build accurate models, the project has provided a roadmap for future development as additional data is collected.

Working with NICD

A key strength of the project was the collaborative approach taken by NICD, combining technical expertise with tailored support to meet Crop Intellect’s needs.

NICD worked closely with the team to ensure that complex data science concepts were accessible and practical, supporting learning alongside delivery and building confidence throughout the engagement.

“I was starting with very limited knowledge… they were very patient and they made it super interesting.” Yusuf Khambhati, Research Scientist, Crop Intellect.

Clear communication and a structured delivery approach ensured the project remained aligned with business objectives while allowing flexibility to explore improvements and future opportunities.
The collaboration provided both immediate value and a strong foundation for continued development, equipping Crop Intellect with the tools and understanding needed to build on the work independently.

This project allowed me and the NICD team to apply our expertise to Crop Intellect’s data and provide valuable insights. When we delivered these insights on Crop Intellect’s data and methodologies, Yusuf and the team were curious, open-minded, and inquisitive. This sparked thought-provoking discussions and gave Crop Intellect interesting avenues of future work to embark on beyond this project.

– Owen Li, Data Scientist at NICD

Looking ahead

With improved data infrastructure and new technical capabilities in place, Crop Intellect is well positioned to build on this work.

As additional field data becomes available in future growing seasons, it will enable the technology to scale through further development of predictive models and continue to make more efficient the validation of the environmental impact of R-leaf.