Linguistic atlases: Many linguistic atlases have been built over the years to represent lexical variation in different areas of Italy. Links to the main atlases can be found in a list of atlases by Accademia della Crusca, although most of them are not actively maintained anymore

Lexicographic resources: Links to a variety of dictionaries can be found in a curated list of dictionaries by Accademia della Crusca

Other data sources: Links to corpora and databases for Italian can be found in the list of language data by Accademia della Crusca

HuggingFace: Huggingface provides a large number of pre-trained language models for many languages in its model hub, as well as thousands of datasets for a variety of tasks in its dataset hub. You can use their python library for designing solutions (tutorial here)

MaChAmp: A toolkit for easy fine-tuning of contextualized embeddings (e.g., from the HuggingFace model hub) in single- and multi-task settings, which supports a large variety of tasks (including classification and regression, useful for GeoLingIt). You can find details on how it works on the description paper, and follow the tutorial here

Varieties of the Boot: A local, multidisciplinary community for the study, documentation, and preservation of languages and dialects of Italy through responsible development of language and speech technologies. Link to the community here