Code and Simulation output for paper "Geographical and temporal variations in stone tools technologies during the transition to farming in prehistoric Japan"

This Zenodo repository contains data and scripts associated with the paper titled:"Geographical and Temporal Variations in Stone Tools Technologies During the Transition to Farming in Prehistoric Japan" by Simon Carrignon, Leah M. Brainerd, Shinya Shoda, Chris Stevens, Glynnis Maynard, Enrico R. Crema The project as three main aspects:1. Creation and cleaning of dataset2. Compute Correspondance Analysis3. Fit Gaussian process and predict missing data 1 and 2 are done by `combineAndclean.R`, but sometime redone locally using `countmatrix`. 3 is done by `compute_gp_ca1.R` Creation, cleaning and CA The script `scripts/combineAndclean.R` allow to recreated the dataset used in the paper from the raw dataset collected by the ENCOUNTER project. The output of the this script is stored in `data/finalsubset.RDS` and is the file that will be then used by the other scripts used to generated the final material of the paper (`scripts/Figure2.R`,`scripts/Figure3.R`,...). The script generates two files: `data/finalsubset.RDS` and `data/countmatrix.RDS`. finalsubset.RDS: This file contains the presence/absence data of all lithics in their original categories, along with additional columns providing detailed metadata: - `sitename`: Site where the phase has been excavated- `address`: Original address of the excavation attributed to the phase- `periods`: Original period attributed to the phase- `year`: Year of publication of the phase- `StartDatePhase`: When available, start of the phase- `EndDatePhase`: When available, end of the phase- `JCODE`: Japanese code for the prefecture- `prefecture`: Prefecture where the site is- `Region`: Higher level region where the prefecture is- `Area`: Rice area following Crema 2023- `side`: East or West?- `wide_periods`: The 4 periods used in the paper- `highlevel_periods`: Jomon or Yayoi?- `lowlevel_periods`: The 4 periods used in the paper- `geometry`: Lat/Lon & CRS (to be read by sf so it can be directly translated into a spatial object)- `jomon_index`: The first dimension of the correspondence analysis- `ca2`: The second dimension of the correspondence analysis countmatrix.RDS: This file stores the original count of the data, which can be converted into a pure integer matrix. File Structure Folder `data/` Data, in CSV, shapefile or RDS forms, used during the analyses: + `data/geo/`: shapefiles and other file related to geographic information+ `data/finalsubset.RDS`: A R version of the data set used for the analyses+ `data/prefecture_region_match.csv`: A lookup table in a CSV from [Crema et al. 2022](https://github.com/ercrema/yayoi_rice_dispersal) that gives which regions falls within wich rice area.+ `data/lithics_periods.csv`: a lookup table that gives the correspondance between each periods from the dataset and the 4 periods used in our analysis + `data/Lithics DB - Categories.csv`: another lookup table that provides the correspondence between lithics and the categories they pertain to.data/prefecture_region_match.csv: a lookup table that provides the correspondence between Japanese prefectures and larger regions.+ `data/simulation/`: `RDS` files storing result of the MCMC: the initial samples for the 4 periods and the predictions. Folder `scripts/` All R files used to clean, prepare and analyses the dataset + `scripts/combineAndclean.R`: A R script used to combine and clean the original dataset into the one used for the analyses + `scripts/extrapol.R`: A R script used to extrapolate value from gaussian process model+ `scripts/compute_gp_ca1.R`: A R script that create and fit the guassian porcess to the data+ `scripts/supplementary_material.R`: A R script that recreatea all figures and tables for supplementary material, including all MCMC summaries+ `scripts/figures/`: All R scripts used to generate the Figures+ `scripts/generateFigures.sh`: use by the docker that recreate the main figures Folder `dockerfiles/` To simplify reproduction of the results we provide Dockerfiles that allow to build image to reproduce the figure in the main paper and in the supplementary material. We use `-v` (`--volume`) option of docker to mount a folder where we can store the results- `dockerfiles/Dockerfile.mainfigures`: to recreate the main figures: + build the image: `docker build -f dockerfiles/Dockerfile.mainfigures -t mainfigure .` + run the image: `docker run -v $PWD/mainfigure:/app/mainfigures mainfigure` After this, results should be in `mainfigure/`- `dockerfiles/Dockerfile.si`: to recreate the main figures: + build the image: `docker build -f dockerfiles/Dockerfile.si -t si .` + run the image: `docker run -e OUTFOLD="si_files" -v $PWD/si_files:/app/si_files si` After this, results should be in `si_files/` (sorry the way to pass the environmental variable that set the output dir is slightly different for the tow dockers, in the firs its in the script `scripts/generateFigures.sh`, in the second it is passe using `-e` Folder `R/` This centralise a few function that are potentially used throughout the project - `R/mergeandgroup.R` this provies the function `st_mergeandgroup` which allow to merge contexts following different criteria- `R/plot_density.R` this provies the function `plot_density`,a wrapper to R base `density` used to generte Figure 3- `R/categories.R` This provides a few functions that are used to handle lithics categories and their translation. It provides a list to go from a Japanese name to its English translation and from a lithics type to the categories it falls within. Funding This work is supported by the European Research Council under the ERC Starting Grant (ENCOUNTER; Project No. 801953).