Tools & Data

Open-source software and replication material

Code and data behind my studies are released on GitHub and the Open Science Framework where copyright allows. Everything below is Python unless stated otherwise.

Software

bitig — computational stylometry

Authorship attribution, author-group style comparison and forensic authorship analysis: Delta, Zeta, PCA/MDS, clustering, bootstrap consensus trees and classification (the core of R’s Stylo), plus a spaCy pipeline, transformer embeddings, a Bayesian layer (PyMC) and likelihood-ratio–calibrated authorship verification. Supports EN, TR, DE, ES, FR. Named after the Old Turkic word for inscription. Documentation · GitHub · PyPI BSD-3-Clause

Linguistic Clusters Framework

Pipeline for identifying stable distributional patterns in learner English passives: instance-level association measures (transitional probability, ΔP, boundary entropy), schema acceptance via NPMI or slot entropy, and validation by permutation, sensitivity and ablation analysis. Replication material for the Corpus Linguistics and Linguistic Theory (2026) article. GitHub · OSF

Metalinguistics — metadiscourse analysis

Detection and classification of metadiscourse markers in academic text (Hyland’s eight categories) with a transformer-based spaCy pipeline and a feature-based classifier; distinguishes academic from conversational uses. Used in the Journal of Language Research (2025) study on Turkish EFL metadiscourse. GitHub MIT

PassivePy — passive voice detection

Passive construction detection and pattern analysis, with comparison of detection methods and visualisation. GitHub

Tense annotation tool

Web application for annotating English tense–aspect structures in learner texts by CEFR level: role-based access, categorised tense options with instructions, auto-save, progress tracking and CSV export. GitHub

Grammar research with spaCy

Collection of reproducible pipelines for grammar analysis in learner English, each module tied to a published study — among them the dative-alternation analysis behind the TAPSLA (2024) article. GitHub

Lexical complexity across CEFR levels

Code base for Bozdağ & Kilimci (2024), Lexical complexity and language proficiency: An investigation of indices across CEFR levels. GitHub

Corpora I work with

ICLE (International Corpus of Learner English, incl. the Turkish sub-corpus TICLE), EFCAMDAT, PELIC and the Turkish National Corpus. I do not redistribute licensed corpus data; scripts expect the user’s own copy.