PODCAST_LEZ: A Feminist Audio-Digital Corpus of French-Language Lesbian Podcasts — Metadata, Acquisition Pipeline, and Analytical Framework (v1.0.0)

PODCAST_LEZ is a curated research corpus of French-language lesbian podcasts, developed within the LZWRLD programme (CNRS, UMR ESO 6590). It is conceived as the empirical foundation for a sustained research reflection on the narrativisation of lesbian and queer trajectories and experiences: how lives are recounted, given shape and made shareable through the podcast as a narrative medium. The dataset provides the metadata infrastructure, acquisition pipeline and analytical framework for studying these lesbian audio-digital narratives as qualitative data in the social sciences. It catalogues French-language lesbian podcasts (28 programmes, 502 episodes tracked) produced in France, Switzerland, Belgium and Quebec between 2018 and 2026. Scientific rationale. The corpus is designed to interrogate how lesbian and queer trajectories and experiences are narrated through an intimate sonic medium, and how this narrativisation produces specific space-times of existence. Lesbian podcasts operate as audio-numeric counterpublics (Fraser 1990; Berlant and Warner 1998) and as technologies of intimacy (Lindgren 2021) in which speakers compose their own spatial and temporal coordinates: the time-spaces of the closet, of coming out, of the first encounter, of chosen family and community, of ageing, of exile and return. Read through narrative identity (Ricoeur 1990; Plummer 1995), queer phenomenology (Ahmed 2006) and queer temporality and chrononormativity (Halberstam 2005; Freeman 2010; Munoz 2009), the corpus documents how told lives deviate from the straight line of heteronormative time and re-orient bodies, objects and territories. It operationalises the concept of lesbotopia (Plard 2026, HAL hal-05601947v1) together with three original analytical propositions: the situated podcast, narrative lesbotopia (an audio-digital space that enables lesbian existence) and friendship infrastructure. It thereby asks how a non-visual, durational, voiced medium narrates and archives spatialities and temporalities that written and visual sources cannot capture, treating the grain of the voice, prosody, silence and laughter as analytical data in their own right. In doing so the corpus serves as a basis for reflection on the narrativisation of lesbian and queer lives, and speaks to current debates on queer and lesbian space-time, on sonic and intermedial aesthetics, and on situated, sensitive methodologies for born-digital qualitative materials. Contents (version 1.0.0-audio, metadata only). The dataset does not include audio files or full transcriptions (copyright preserved). It provides: (1) a master metadata database (BDD_CORPUS_MASTER.xlsx) cataloguing the podcasts and tracking episodes by status, typology and relevance/narrativity scoring; (2) a reproducible acquisition pipeline in Python (RSS resolution, download, conversion, database integration), transposed from the six-layer STRIDE pipeline; (3) a D1-D6 analytical codebook for feminist qualitative coding (Corporeality, Resonance, Temporality, Spatiality, Affects, Voice/Agency); (4) the methodology, FAIR data-management plan and data-paper document; (5) verified bibliographic references. Ethics and law. RGPD compliant; third parties pseudonymised. Original audio is not redistributed: analyses rely on the text-and-data-mining research exception (Art. L.122-5-3 CPI; EU Directive 2019/790, Art. 3). Files are under embargo until 1 January 2028 to secure scholarly priority during corpus completion and data-paper submission. Future versions will add Whisper transcriptions via TGIR Huma-Num (version 2.0.0-text) and STRIDE enrichment schemas (version 3.0.0-enriched).