An easy way to examine archaeological count data. This package provides a convenient and reproducible toolkit for relative and absolute dating and analysis of (chronological) patterns. It includes functions for matrix seriation (reciprocal ranking, CA-based seriation), chronological modeling and dating of archaeological assemblages and/or objects. Beyond these, the package provides several tests and measures of diversity: heterogeneity and evenness (Brillouin, Shannon, Simpson, etc.), richness and rarefaction (Chao1, Chao2, ACE, ICE, etc.), turnover and similarity (Brainerd-Robinson, etc.). The package make it easy to visualize count data and statistical thresholds: rank vs. abundance plots, heatmaps, Ford (1962) and Bertin (1977) diagrams.
To cite tabula in publications please use:
Frerebeau, N. (2019). tabula: An R Package for Analysis, Seriation, and Visualization of Archaeological Count Data. Journal of Open Source Software, 4(44), 1821. DOI 10.21105/joss.01821.
You can install the released version of tabula from CRAN with:
Or install the development version from GitHub with:
tabula provides a set of S4 classes that extend the basic
matrix data type. These new classes represent different special types of matrix.
CountMatrixrepresents count data,
FrequencyMatrixrepresents relative frequency data.
IncidenceMatrixrepresents presence/absence data.
OccurrenceMatrixrepresents a co-occurrence matrix.
SimilarityMatrixrepresents a (dis)similarity matrix.
It assumes that you keep your data tidy: each variable (type/taxa) must be saved in its own column and each observation (sample/case) must be saved in its own row.
These new classes are of simple use, on the same way as the base
tabula uses coercing mechanisms (with validation methods) for data type conversions:
## Create a count matrix A1 <- CountMatrix(data = sample(0:10, 100, TRUE), nrow = 10, ncol = 10) ## Coerce counts to frequencies B <- as_frequency(A1) ## Row sums are internally stored before coercing to a frequency matrix ## (use totals() to get these values) ## This allows to restore the source data A2 <- as_count(B) all(A1 == A2) #>  TRUE ## Coerce to presence/absence C <- as_incidence(A1) ## Coerce to a co-occurrence matrix D <- as_occurrence(A1)
Several types of graphs are available in tabula which uses ggplot2 for plotting informations. This makes it easy to customize diagrams (e.g. using themes and scales).
Spot matrix allows direct examination of data:
# Plot co-occurrence of types # (i.e. how many times (percent) each pairs of taxa occur together # in at least one sample.) mississippi %>% as_occurrence() %>% plot_spot() + ggplot2::labs(size = "", colour = "Co-occurrence") + ggplot2::theme(legend.box = "horizontal") + khroma::scale_colour_YlOrBr()
Bertin or Ford (battleship curve) diagrams can be plotted, with statistic threshold (including B. Desachy’s sériographe).
# Build an incidence matrix with random data set.seed(12345) incidence <- IncidenceMatrix(data = sample(0:1, 400, TRUE, c(0.6, 0.4)), nrow = 20) # Get seriation order on rows and columns # Correspondance analysis-based seriation (indices <- seriate_reciprocal(incidence, margin = c(1, 2))) #> Permutation order for matrix seriation: #> Matrix ID: 2b312a8d-5cc5-491c-ac5d-37eecd2f9bf2 #> Row order: 1 4 20 3 9 16 19 10 13 2 11 7 17 5 6 18 14 15 8 12 #> Column order: 1 16 9 4 8 14 3 20 13 2 6 18 7 17 5 11 19 12 15 10 #> Method: reciprocal # Permute matrix rows and columns incidence2 <- permute(incidence, indices)
# Plot matrix plot_heatmap(incidence) + ggplot2::labs(title = "Original matrix") + ggplot2::scale_fill_manual(values = c("TRUE" = "black", "FALSE" = "white")) plot_heatmap(incidence2) + ggplot2::labs(title = "Rearranged matrix") + ggplot2::scale_fill_manual(values = c("TRUE" = "black", "FALSE" = "white"))
This package provides an implementation of the chronological modeling method developed by Bellanger and Husi (2012). This method is slightly modified here and allows the construction of different probability density curves of archaeological assemblage dates (event, activity and tempo). Note that this implementation is experimental (see
# Coerce dataset to abundance (count) matrix zuni_counts <- as_count(zuni) # Assume that some assemblages are reliably dated (this is NOT a real example) # The names of the vector entries must match the names of the assemblages set_dates(zuni_counts) <- c( LZ0569 = 1097, LZ0279 = 1119, CS16 = 1328, LZ0066 = 1111, LZ0852 = 1216, LZ1209 = 1251, CS144 = 1262, LZ0563 = 1206, LZ0329 = 1076, LZ0005Q = 859, LZ0322 = 1109, LZ0067 = 863, LZ0578 = 1180, LZ0227 = 1104, LZ0610 = 1074 ) # Model the event date for each assemblage model <- date_event(zuni_counts, cutoff = 90) # Plot activity and tempo distributions plot_date(model, type = "activity", select = "LZ1105") + ggplot2::labs(title = "Activity plot") + ggplot2::theme_bw() plot_date(model, type = "tempo", select = "LZ1105") + ggplot2::labs(title = "Tempo plot") + ggplot2::theme_bw()
Diversity can be measured according to several indices (sometimes referred to as indices of heterogeneity):
mississippi %>% as_count() %>% index_heterogeneity(method = "shannon") #> <HeterogeneityIndex: shannon> #> index #> 10-P-1 1.2027955 #> 11-N-9 0.7646565 #> 11-N-1 0.9293974 #> 11-O-10 0.8228576 #> 11-N-4 0.7901428 #> 13-N-5 0.9998430 #> 13-N-4 1.2051989 #> 13-N-16 1.1776226 #> 13-O-11 1.1533432 #> 13-O-10 1.2884172 #> 13-P-1 1.1725355 #> 13-P-8 1.5296294 #> 13-P-10 1.7952443 #> 13-O-7 1.1627477 #> 13-O-5 1.0718463 #> 13-N-21 0.9205717 #> 12-O-5 1.1751002 #> Holden Lake 0.7307620 #> 13-N-15 1.1270126 #> 12-N-3 1.0270291 ## Test difference in Shannon diversity between assemblages ## (returns a matrix of adjusted p values) mississippi[1:5, ] %>% as_count() %>% test_diversity() #> 10-P-1 11-N-9 11-N-1 11-O-10 #> 11-N-9 0.000000e+00 NA NA NA #> 11-N-1 3.609626e-08 8.538298e-05 NA NA #> 11-O-10 2.415845e-13 4.735511e-01 2.860461e-02 NA #> 11-N-4 0.000000e+00 7.116363e-01 7.961107e-05 0.7116363
simpson methods return a dominance index, not the reciprocal form usually adopted, so that an increase in the value of the index accompanies a decrease in diversity.
Corresponding evenness (i.e. a measure of how evenly individuals are distributed across the sample) can also be computed, as well as richness and rarefaction.
Several methods can be used to ascertain the degree of turnover in taxa composition along a gradient on qualitative (presence/absence) data. It assumes that the order of the matrix rows (from 1 to n) follows the progression along the gradient/transect.
Diversity can also be measured by addressing similarity between pairs of sites:
The Frequency Increment Test can be used to assess the detection and quantification of selective processes in the archaeological record.
## Keep only decoration types that have a maximum frequency of at least 50 keep <- apply(X = merzbach, MARGIN = 2, FUN = function(x) max(x) >= 50) merzbach_count <- as_count(merzbach[, keep]) ## The data are grouped by phase ## We use the row names as time coordinates (roman numerals) set_dates(merzbach_count) <- rownames(merzbach) ## Plot time vs abundance and highlight selection plot_time(merzbach_count, highlight = "FIT", roll = TRUE) + ggplot2::theme_bw() + khroma::scale_color_contrast()