We present ConsumerBR, a large-scale corpus of consumer complaints and company responses in Brazilian Portuguese, compiled from publicly available data on the Consumidor.gov.br platform. The corpus comprises over 3.1 million consumer–company interactions collected between 2021 and 2025 and combines anonymized textual content with rich structured metadata, including temporal information, complaint outcomes, and consumer satisfaction indicators. We describe a data collection strategy tailored to the platform’s dynamic interface, a preprocessing pipeline that includes response clustering to identify template-based replies, and a hybrid anonymization approach designed to mitigate privacy risks. We also provide a detailed statistical characterization of the corpus, highlighting its scale, coverage, and distributional properties. ConsumerBR is publicly available for research purposes and supports a wide range of applications, including complaint analysis, sentiment modeling, dialogue and response generation, and preference-based evaluation.

Dense retrieval is a critical component of Retrieval-Augmented Generation (RAG) systems and is highly sensitive to document representations. In consumer complaint settings, raw interaction texts are often lengthy and noisy, which limits retrieval effectiveness. This paper investigates whether schema-guided structured summaries can improve dense retrieval in RAG. We compare embeddings derived from raw interaction texts and from LLM-generated structured summaries in a controlled evaluation on Portuguese-language consumer complaints. Summary-based retrieval achieves a Recall@1 of 0.527, compared to 0.001 when indexing raw interactions, and reaches Recall@10 of 0.610, demonstrating gains of more than two orders of magnitude. These results show that structured summaries enable more effective and reliable retrieval at low cutoffs, making them particularly suitable for RAG pipelines.

In public procurement, establishing reference prices is essential to guide competitors in setting product prices. Group-purchased products, which are not standardized by default, are necessary to estimate reference prices. Text clustering techniques can be used to group similar items based on their descriptions, enabling the definition of reference prices for specific products or services. However, selecting an appropriate representation for text is challenging. This paper introduces a framework for text cleaning, extraction, and representation. We test eight distinct sentence representations tailored for public procurement item descriptions. Among these representations, we propose an approach that captures the most important components of item descriptions. Through extensive evaluation of a dataset comprising over 2 million items, our findings show that using sophisticated supervised methods to derive vectors for unsupervised tasks offers little advantages over leveraging unsupervised methods. Our results also highlight that domain-specific contextual knowledge is crucial for representation improvement.

The intersection of natural language processing (NLP) and literary analysis has yielded valuable insights and applications across various languages. However, the scarcity of labeled data tailored for Portuguese literary texts poses a notable challenge. To address this gap, we present the PPORTAL_ner corpus, an annotated dataset that simplifies the development of Named Entity Recognition (NER) models specifically adapted for Portuguese literary works. Our corpus includes annotations of PER, LOC, GPE, ORG, and DATE entities within a diverse set of 25 literary texts. Annotation of the corpus involved a two-step process: initial pre-annotation using a pre-trained spaCy model followed by correction and refinement using the Prodigy annotation tool. With a total of 125,059 tokens and 5,266 annotated entities, PPORTAL_ner corpus significantly enriches the landscape of resources available for computational literary analysis in Portuguese. This paper details the annotation methodology, guidelines, and dataset statistics while also evaluating four NER models over the PPORTAL_ner corpus. Our evaluation analysis reveals that fine-tuning on domain-specific data significantly improves NER model performance, demonstrating the value of the PPORTAL_ner corpus for developing domain-specific language models.