digital-humanities-guide

Digital Humanities Guide

A skill for applying computational and quantitative methods to humanities research. Covers text mining, network analysis, spatial humanities, and digital archival methods. Designed for researchers bridging traditional humanities with data-driven approaches.

Text Mining and Distant Reading

Corpus Preparation

import re
from collections import Counter

def prepare_corpus(texts: list[str], stopwords: set = None) -> list[list[str]]:
    """
    Tokenize and clean a corpus of texts for analysis.

    Args:
        texts: List of raw text strings
        stopwords: Set of words to remove
    Returns:
        List of tokenized, cleaned documents
    """
    if stopwords is None:
        stopwords = {'the', 'a', 'an', 'and', 'or', 'but', 'in', 'on',
                     'at', 'to', 'for', 'of', 'with', 'is', 'was', 'are'}

    processed = []
    for text in texts:
        # Lowercase and remove punctuation
        tokens = re.findall(r'\b[a-z]+\b', text.lower())
        # Remove stopwords and short tokens
        tokens = [t for t in tokens if t not in stopwords and len(t) > 2]
        processed.append(tokens)
    return processed

def compute_tfidf(corpus: list[list[str]]) -> dict:
    """Compute TF-IDF scores for term importance analysis."""
    import math
    n_docs = len(corpus)
    # Document frequency
    df = Counter()
    for doc in corpus:
        df.update(set(doc))
    # TF-IDF per document
    tfidf_scores = []
    for doc in corpus:
        tf = Counter(doc)
        total = len(doc)
        scores = {}
        for term, count in tf.items():
            tf_val = count / total
            idf_val = math.log(n_docs / (1 + df[term]))
            scores[term] = tf_val * idf_val
        tfidf_scores.append(scores)
    return tfidf_scores

Topic Modeling

Apply Latent Dirichlet Allocation (LDA) to discover thematic structures in large text corpora:

from gensim import corpora, models

def run_topic_model(corpus: list[list[str]], n_topics: int = 10,
                     passes: int = 15) -> models.LdaModel:
    """
    Train an LDA topic model on a preprocessed corpus.
    """
    dictionary = corpora.Dictionary(corpus)
    dictionary.filter_extremes(no_below=5, no_above=0.5)
    bow_corpus = [dictionary.doc2bow(doc) for doc in corpus]

    lda_model = models.LdaModel(
        bow_corpus,
        num_topics=n_topics,
        id2word=dictionary,
        passes=passes,
        random_state=42,
        alpha='auto',
        eta='auto'
    )
    return lda_model

# Print top words per topic
# for idx, topic in lda_model.print_topics(-1):
#     print(f"Topic {idx}: {topic}")