University of Amsterdam

Overview

At Palmyra, we believe technology can transform supply chains while empowering local communities. This belief is at the core of our work with Nature’s Nectar in Zambia, a sustainable beekeeping initiative focused on transparency, regulatory compliance, and measurable social impact.

Powered by Palmyra Pro, our traceability and compliance platform, the project captures who produces, harvests, and processes honey—ensuring authenticity, transparency, and trust while supporting economic opportunities for local beekeepers.

In 2025, we took this initiative further through a formal collaboration with BSc Business Analytics students from the University of Amsterdam, launching a real-world Data Challenge built on Palmyra Pro traceability data.

From Traceability to Intelligence

This collaboration went beyond academic research. It was designed as a real-world analytics initiative to extract deeper insights from live supply-chain data and strengthen sustainable forest management and honey production.

Using a reverse-mentoring approach, students applied advanced analytical methods to Palmyra Pro datasets—moving beyond traditional ESG reporting to focus on risk prediction, optimisation, and decision-support intelligence.

Their fresh perspectives helped uncover new ways to optimise operations, scale impact, and improve resilience across the supply chain.

Analytical Methods & Insights

The project applied a broad set of advanced analytical techniques, including:

• Spatial and correlation analyses combining forest loss data with beehive placement to assess the relationship between forest degradation and beekeeping activity.

• Random Forest Regression with SHAP (Shapley Additive Explanations) to identify the most important variables influencing honey yield.

• Grid-based spatial analysis to highlight unexplored zones suitable for hive expansion and onboarding new farmers.

• SARIMAX time-series models to forecast climate effects on honey production, achieving 73–90% accuracy.

• Regularised regression models (Ridge & Lasso) to identify significant climate variables and early warning signals for adaptive management (≈90% accuracy).

• Spline regression and clustering models applied to demographic and transactional data, predicting yields with ~60% test accuracy and segmenting farmers into four distinct producer profiles.

Impact & Outcomes

• Improved understanding of how climate, forest conditions, and hive placement influence honey yield

• Data-driven guidance for sustainable hive expansion and farmer onboarding

• Enhanced decision-making tools for adaptive forest and production management

• Demonstrated the power of Palmyra Pro data beyond compliance—into predictive and optimisation intelligence

Most importantly, the collaboration reinforced the value of working with motivated, skilled students:
the results were so strong that Palmyra ultimately hired one of the project participants.

Project Details

• Collaboration Partner: Palmyra

• Academic Partner: University of Amsterdam

• Programme: BSc Business Analytics

• Year: 2025

• Project Members:
  Deniz Cayci · Gergana Ivanova · My Ngoc Nguyen · Abhinav Bharat Prem