Memecry:Tracing the Repetition-with-Variation of Formulas on 4chan/pol/

Datasets underlying the analysis of the paper "Memecry: Tracing the Repetition-with-Variation of Formulas on 4chan/pol/

This upload includes the following:

seedwords.csv: A .csv file with terms we used as a seed list to filter for 4chan/pol/-post containing vernacular.

seedword-network_x.gdf/gephi: .gdf and .gephi network files for NPMI-weighted co-word networks of /pol/-posts. We only included posts that contained one of the aforementioned seed list words.

twoflow-data_x.xlsx: .xlsx files with data on triplets common to 4chan/pol/. We identified these three-word sequences through the above network files. For example: "gr8 b8 m8", "orange man bad", "lurk moar newfag". The Excel data on these triplet includes:

The absolute amount of /pol/-posts per year mentioning the triplets (within a window of five words).

The average NPMI scores between the three triplet words per year.

The top co-words per year having an average NPMI higher than 0.18 with two of the three triplet words.

triplets.csv: A .csv file with the extracted triplets, including their common appearance as memetic phrases and a short explanation.

This data was used for "two-flow graphs" available at oilab.eu/formulas/.

See the paper for full explanations on the data.