DH Tools & Software
Open-source computational tools for Digital Humanities research. Built for reproducible literary analysis, trauma narrative detection, and cross-cultural NLP. Free to use, fork, and extend.
Reproducibility First
Every tool includes configuration files, seed parameters, and version-locked dependencies. Research results should be independently verifiable.
Transparent Methodology
Detailed documentation of model choices, training data provenance, and known limitations. No black boxes.
Interdisciplinary Design
Built for humanists, not just engineers. CLI and GUI interfaces, human-readable output formats, and extensive inline documentation.
LitKit
Literary Analysis Toolkit
A modular NLP pipeline for computational literary studies. LitKit performs sentiment analysis, thematic clustering, narrative arc detection, and stylometric fingerprinting on literary corpora. Designed for scholars who need reproducible, configurable text analysis without writing custom pipelines from scratch.
Capabilities
- ▸Sentiment trajectory mapping across narrative arcs
- ▸TF-IDF and LDA topic modeling for thematic analysis
- ▸Stylometric comparison across authors and periods
- ▸Named Entity Recognition with literary entity types
- ▸Exportable visualizations (PNG, SVG, CSV)
Trauma-Linguistics-Engine
Affective Computing for Trauma Narratives
A specialized NLP engine for detecting trauma markers in literary and clinical texts. Uses fine-tuned transformer models (BERT, RoBERTa) to identify linguistic patterns associated with PTSD, dissociation, fragmented memory, and temporal distortion. Built for interdisciplinary researchers bridging psychology and literary studies.
Capabilities
- ▸Fine-tuned BERT classifier for trauma-adjacent language
- ▸Temporal distortion detection in narrative sequences
- ▸Dissociative language pattern recognition
- ▸Comparative analysis across cultural/linguistic traditions
- ▸Integration with standard clinical PTSD assessment markers
Corpus Explorer
Visual Corpus Navigation
An interactive visualization tool for exploring large literary corpora. Generates zoomable topic maps, word frequency heatmaps, and inter-text relationship graphs. Useful for distant reading projects where manual close reading is impractical at scale.
Capabilities
- ▸Interactive word frequency heatmaps
- ▸T-SNE and UMAP dimensionality reduction for text clustering
- ▸Collocation network visualization
- ▸Concordance search with keyword-in-context (KWIC) display
- ▸Batch processing for corpora with 1000+ documents
Dysfluency Analyzer
Computational Stammering Detection
A targeted NLP module for identifying and categorizing speech dysfluency patterns in transcribed text. Detects repetitions, prolongations, blocks, interjections, and broken words. Designed for disability studies researchers analyzing representations of stammering in literature and film subtitles.
Capabilities
- ▸Dysfluency type classification (repetition, prolongation, block)
- ▸Frequency and distribution analysis across texts
- ▸Comparison with fluent baseline corpora
- ▸Support for Bengali, Hindi, and English transcripts
- ▸Statistical significance testing for dysfluency density
PostcolonialNLP
Decolonizing Text Analysis Pipelines
A framework for adapting standard NLP tools to non-Western literary traditions. Addresses the inherent biases in English-centric NLP models by providing custom tokenizers, stopword lists, and sentiment lexicons for Bengali and Hindi literary texts. Includes critical documentation on how standard NLP tools fail with postcolonial literatures.
Capabilities
- ▸Custom tokenizers for Bengali and Hindi literary texts
- ▸Culturally-calibrated sentiment lexicons
- ▸Postcolonial-aware Named Entity Recognition
- ▸Bias audit tools for standard NLP pipelines
- ▸Documentation of failure modes in colonial-language NLP
Use These Tools? Cite Them.
If these tools are useful in your research, please cite them in your publications. Open-source DH tools grow through academic citation networks.
Share these tools:
Subscribe to Newsletter
Get the latest posts delivered right to your inbox
Join 1,000+ readers. No spam, unsubscribe anytime.