Welcome to the repository for the Querelle des Colleges (QdC).
It contains dynamic network visualisations from two publications:
Tidman G. (2023). The Emergence of Literature in Eighteenth-Century France: The Battle of the School Books. Liverpool: Liverpool University Press.
Authors (Submitted). Movers and Haters. Dynamic Network Visualisation and Analysis of an Historical French Controversy
ARTICLE
The dynamic visualisations prepared for Tidman (2023) can be found within the repo's docs directory.
All other files and directories relate to the analyses carried out for Authors (Submitted):
- article_figures_and_appendices directory: contains outputs (figures and appendices) used in Authors (Submitted).
- original_qdc_datasets directory: contains versions of the original dataset used to create all QdC statistics and visuals. Files are dated, with dates contained in file names; all current statistics and visuals for Authors (Submitted) were prepared from QDC_2024_06_07.xlsx.
- dynamic_network_visuals directory: contains three dynamic network visualisations each from the "text" and "person" networks of the QdC. Three visualisations were created for each network as, while all visualisations were created using the same procedures (in particular, the Kamada-Kawai force-directed algorithm for determining node position), those procedures can result in different node positions; The three visuals therefore serve to mitigate against biases that could occur from overly relying on single visualisations.
- community_network_visuals directory: contains the static network visualisations mentioned in the "Community" subsection of the Results section of Authors (Submitted).
- charts directory: contains line graphs and histograms from the "centrality" and "community" subsections (Results section) of the article
- stats directory: contains statistics from the person and text networks of the QdC that are reported in the article (in each subsection of the Results section)
All network analyses and visualisations for the article were produced using R version 4.1.3. A table of packages and package versions used for the analyses can be found below.
The "main.R" file, contained within the repo's root directory, can be used to reproduce the article analyses, as well as the dynamic visualisations. Each analysis or visualisation can be produced by setting the correct configurable values in lines 1-32 of the file, setting the relevant object value to TRUE, and running the entire file. Outputs are exported to the repo's directories as specified above. Guidance on how to set each configurable value is included within the "main.R" file.
For example, to produce the dynamic visualisations, the produce_dynamic_visual object should be set to TRUE (i.e. the beginning of line XX should be "produce_dynamic_visual <- TRUE"). the text_or_pers object in line XX should be set to "text" to produce the dynamic text network visual and to "pers" to produce the dynamic person network visual. The code for producing the dynamic text and person visuals can respectively be found in the "produce_dynamic_text_network.R" and "produce_dynamic_pers_network.R" files.
Other article analyses, listed below according to article section, can be run by setting the following objects to true:
- Table 1: "run_ad_hoc_analyses" should be set to TRUE. Results - for the text or person network, according to the value of the text_or_pers object (see above) - are stored in the "table_1_results" object. Code can be found in the "ad_hoc_analyses.R" file
- Figures 7, 8: can be produced by setting "produce_centrality_histogram" to TRUE, and setting "histogram_measure" to "degree". Figures are produced by the "produce_centrality_histogram.R" file.
- Statistics on ties sent/received by various actors (lines 323-325, 370-371): produced by setting "produce_statistics_by_slice_or_year" to TRUE. This runs the "QDC_stats_by_slice.R" file, which produces .xlsx files containing various network statistics by slice or year of the QdC, either for the person or text network. The files are produced in the "stats" directory of the repository. The statistics themselves can then be retrieved from the .xlsx files, in the files' "degree" and "degree_neg" sheets, which show degree values per year or slice of the Querelle (to find the final degree values over the entire querelle, go to the rightmost columns in those sheets).
- Statistics on texts published in 1762 and 1763 and their references (lines 371-373): produced by setting "run_ad_hoc_analyses" to TRUE. The "references_of_texts_62_63" object stores the number of references received by texts from authors publishing in 1762 and 1763 according to the quality of the reference. Find the row for Rousseau (1762), and add up the columns for negative (-2 and -1) and ambivalent references (3), to find the number of texts who send negative/ambivalent references to Rousseau (1762). The "percent_refs_to_rousseau" object stores the proportion of texts published in 1762-1763 who send references to Rousseau.
- Statistics on texts referring to Rousseau (1762) and La Chalotais (1763), and on number of texts making references but receiving no references (lines 409-416): produced by setting "run_ad_hoc_analyses" to TRUE. The indegrees (number of references received) and outdegrees (number of references made) of texts referring to Rousseau and La Chalotais are stored in the "indegree_texts_referring_to_rousseau", "indegree_texts_referring_to_la_chalotais", "outdegree_texts_referring_to_rousseau", and "outdegree_texts_referring_to_la_chalotais". A tabulation of indegrees of texts that make references is stored in the "indegree_all_texts_making_references" object.
- Statistics on texts receiving no references: obtained from the .xlsx files produced by setting "produce_statistics_by_slice_or_year" to TRUE. See the "indegree" sheet within the outputted .xlsx files.
- Network visualisations resulting from the Louvain and Leiden community detection algorithms (Figures 9-12, Appendices F.1-F.3): Produced by setting "produce_community_visuals" to TRUE. Uses the Louvain or Leiden algorithms depending on the configurable "cd_algorithm" value set within main.R. Runs the "produce_communities.R" and "produce_community_visuals.R" files.
- Modularity statistics (Appendices C-D) and charts (Figure 13, Appendix E) resulting from running the Louvain and Leiden algorithms: Produced by setting "produce_modularity_stats" to TRUE. Uses the Louvain or Leiden algorithms depending on the configurable "cd_algorithm" value set within main.R. Runs the "produce_communities.R" and "produce_modularity_stats.R" files.
- Co-references between La Chalotais, Rousseau, D'Alembert, and De l'Education Publique: produced by setting "run_ad_hoc_analyses" to TRUE. The "persons_referring_to_la_chalotais_and_others" object stores the names of persons who refer to Chalotais and to one of the other actors, and how many of the other actors they refer to. "n_persons_referring_to_la_chalotais_and_others" records how many of these persons there are.
- Statistics on Borrelly's undirected eigenvector centrality: obtained from the .xlsx files produced by "QDC_stats_by_slice.R". See the "eigenvector_undirected" sheet within the outputted file(s).
- Borrelly's intervention leading to a drop in modularity: See 'Modularity statistics and charts' mentioned above.
- Statistics on D'Alembert's degree centrality: obtained from the .xlsx files produced by "QDC_stats_by_slice.R". See the "degree" sheet within the outputted file(s).
- Analysis on balance in the ties sent to D'Alembert and Leroy: obtained from the dynamic visualisations.
R Version 4.1.3, as well as the following packages and package versions, were used to produce the article analyses:
| package | version number |
|---|---|
| base R packages (base, stats etc.) | 4.1.3 |
| abind | 1.4-5 |
| Cairo | 1.6-0 |
| data.table | 1.14.8 |
| dplyr | 1.1.2 |
| extrafont | 0.19 |
| ggplot2 | 3.4.2 |
| igraph | 1.4.2 |
| intergraph | 2.0-2 |
| ndtv | 0.13.3 |
| network | 1.18.1 |
| networkDynamic | 0.11.3 |
| rio | 0.5.29 |
| sna | 2.7-1 |
| statnet.common | 4.8.0 |
| tsna | 0.3.5 |