In this folder we present the supplementary materials for our paper: At the Intersection of NLP and Sustainable Development: Exploring the Impact of Demographic-Aware Text Representations in Modeling Value on a Corpus of Interviews.
Here, we present the Demographic-Rich Qualitative UPV-Interviews Dataset (DR-QI) and release the code we used to implement our models.

The DR-QI dataset contains 5,333 sentences. The sentences constitute excerpts from 214 interviews, conducted in 56 rural villages in India and 68 villages in Uganda. The interviews were part of a larger project to develop an impact framework for off-grid energy appliances in Low- and Middle-Income countries (LMICs). The sentences in the DR-QI dataset were annotated with the User-Perceived Value (UPV) approach. The dataset contains additional demographic information for each speaker. The dataset releases the 5,333 translated sentences, along with their UPV annotations. For each speaker, ten self-reported categorical demographic features are included.

We release the DR-QI dataset under a Academic Free License agreement. In order to get access to the DR-QI dataset, please fill out the agreement form in the DR-QI folder and send to j.g.vanboven@students.uu.nl .

The folder DR-QI contains a dataset sample, along with a data statement and an overview of the UPV definitions.
The folder Code presents the code we used to train our models.