Digital humanities applies computing to the study of culture: texts, art, music, maps, archives, languages. It lets a scholar read a million pages at once, link scattered records, or preserve a fragile manuscript as an image. The best work keeps the humanist's questions and caution about interpretation, and uses the machine to scale, not to replace, close reading.
Key concepts
- Text encoding: marking up a text with structure and meaning in a standard such as TEI, in XML.
- Distant reading: counting and modelling words across large collections to see patterns no one could read through.
- Linked data and metadata: describing objects with shared vocabularies so collections can be joined and searched.
- Mapping and networks: places, people and relations as maps and graphs, and what they do and do not show.
- Digitisation and preservation: scanning, optical character recognition, file formats and long-term storage.
- Data critique: every dataset was assembled by someone, with gaps, biases and choices that shape the result.
Vocabulary
- Corpus: a collection of texts chosen for study.
- OCR: optical character recognition, turning a scanned page into text, with errors to check.
- TEI: the Text Encoding Initiative, a community standard for encoding texts.
- Topic model: a statistical method that groups words that tend to occur together in documents.
What to study, in order
- Learn a scripting language and basic text handling: Unicode, regular expressions, XML.
- Encode a short text, a poem or a letter, in TEI and validate it.
- Take a public-domain corpus, count words, compare authors or periods, and question what the counts mean.
- Build a small catalogue with shared metadata and publish it as linked data.
- Make a map or network from historical records, and write down its limits.
Related tags
#nlp #information-retrieval #knowledge-graphs #databases #history