LCAi: Life Cycle Assessment with big data fusion and retrieval-augmented generation -assisted interpretation

This repository accompanies the manuscript “LCAi: Life Cycle Assessment with big data fusion and retrieval-augmented generation-assisted interpretation” and provides the code and data-format documentation for the LCAi proof-of-concept framework. LCAi explores how life cycle assessment (LCA) outputs can inform implementation-oriented pathway mapping through retrieval-augmented generation (RAG). The workflow combines a case-specific scenario anchor, perspective-conditioned retrieval and inference, and a final synthesis of the generated outputs without additional retrieval. The framework uses four complementary evidence perspectives: academic material from Scopus, business descriptions from LinkedIn, public posts and comments from YouTube, Reddit, and Bluesky, and EU-funded project information from CORDIS. The manuscript demonstrates the workflow by exploring a hypothetical hydrogen implementation pathway associated with a diesel-reduction target in Italian apple production and distribution. This application illustrates the workflow’s functionality; it does not establish the environmental superiority or implementation feasibility of hydrogen substitution. The repository contains: The RAG LCAi application code. Complementary documentation illustrating the source-data fields and the perspective-specific text and metadata structures, with one actual example record from each platform. The full collected corpora are not redistributed. The examples document the data used in the study and their structure; they are not sufficient to reproduce the reported empirical outputs. Users seeking to apply the workflow to other contexts must supply suitable datasets and any credentials required by the code.