D-PLACE lets you visualize cross-cultural data in three ways: as a list in a table, on a global map, or on a linguistic tree (i.e., phylogeny or classification). Data points (i.e., rows in the table, points on the map, points on the tree) represent the expression of a particular cultural feature by a particular society at a given time and place, as recorded by the Source. Data points are colour-coded to allow you to quickly assess the diversity of a cultural feature across societies in the table, map, or linguistic tree.
We suggest you start exploring D-PLACE via a search on the Variables page. ‘Variables’ in D-PLACE are cultural features or environmental parameters.
For example, to explore cross-cultural variation in “Games”, start on the Variables page by typing “games” into the text box at the top of the first column.
This should narrow the list of selectable variables from multiple thousand to a small handful. Each variable you see listed in the Variables search table was contributed as part of a particular cross-cultural Dataset and has been coded for a particular Society Set.
Variables are listed by ‘name’ and unique ‘variable ID’ (e.g., “Games [EA035]”, “Games at Initial Funeral Ceremonies [SCCS1987]”), and can be sorted by “Dataset”.
Select “Games [EA035]” by clicking on it (note that the variable ID “EA035” tells you this variable is from the Ethnographic Atlas (EA) dataset – see the Datasets page for more).
You should now find yourself on a results page for the “Games [EA035]” variable. You will see its name and ID repeated at the top of the page, together with a definition that provides more information on how the variable was defined. In the orange box you will see the list of “possible codes” for the variable, as well as the colours that will be used to symbolise each code in the table, map and tree.
You can now switch between a table view, map view and tree view of the coded variable by clicking on the ‘table’, ‘map’ or ‘tree’ icons. NOTE that to view your results on a tree, you must first choose the language family and/or particular phylogeny for which you would like a tree to be produced. You can explore available trees on the Phylogenies page.
Note that you can also add additional variables to your selection at this point. Combining variables allows you to examine co-occurrence patterns for different cultural features. For example, you might be interested in how game type is related to a society’s dependence on agriculture (EA005). In this case, you could start typing “agriculture” into the ‘Combine’ box, and select EA005 from the results that are offered to you. Tip: to maximize the number of societies from your first selection that will be covered by the second selection, make sure the second variable you select is from the same base Dataset (in this case, from the EA).
D-PLACE relies on existing cross-cultural, linguistic and environmental datasets that themselves are the product of decades of work by scientists, translators, transcribers and research participants. The D-PLACE database itself was designed and developed by a team of scientists from a range of disciplines.
D-PLACE was developed with generous support from the National Evolutionary Synthesis Center} and the Max Planck Institute for the Science of Human History.
We use the term “society” to refer to cultural groups in the database. In most cases, a society can be understood to represent a group of people at a focal location with a shared language that differs from that of their neighbors. However, in some cases multiple societies share a language. There is also some variation among authors of different datasets in how societies are delineated, with the same cultural group embedded in a larger unit in one cross-cultural sample, but split into multiple groups in another. For example, the society Murdock (1967) refers to as Tunava includes both the Deep Springs Valley and Fish Lake Valley Paiute groups, whereas Binford (2001) describes the Fish Lake and Deep Springs Paiute as distinct societies. D-PLACE highlights potential links among such societies by assigning them a matched “cross-dataset id” (xd_id), but leaves decisions on when and how to combine data to the user. Each society in D-PLACE must be accompanied by information on its geographic location (latitude and longitude coordinates), main year of documentation, language spoken, and primary sources (e.g., full citations for the ethnographic sources used to code cultural practices).
Variables are cultural features or practices, or environmental descriptors. Variables must be clearly defined (e.g., for a variable “Dependence on fishing”, what is meant by “dependence” (caloric contribution of fishing to the diet?, time spent on fishing relative to other subsistence activities?), and what is meant by “fishing” (does this include shellfish? aquatic mammals?). If categorical, variables must be accompanied by a list of possible ‘codes’ (e.g., variable: Marriage system; possible codes: Monogamous, Polygamous, Polyandrous). Variables can also be continuous (e.g., Total population, Annual rainfall) or ordinal (e.g., Relative settlement size, on a scale of 1-10).
D-PLACE makes accessible existing, coded cultural datasets, and links them with information on environment, language and geography. Therefore, codes are determined before a dataset is imported into D-PLACE. In the case of the ‘core’ D-PLACE datasets, codes were defined by the authors of those datasets (e.g., Ethnographic Atlas codes by George P. Murdock, Binford Hunter-Gatherer codes by Lewis Binford (though, note that in some cases Binford adopted Murdock’s codes).
A dataset is made up of one or more variables that have been coded for one or more societies (aka, a “society set”). See also the FAQ “I would like to contribute data to D-PLACE. Where do I start?” for examples on how we define society sets and datasets.
Each datapoint in D-PLACE is tagged with information on the geographic location (latitude and longitude coordinates), year(“focal year”), language of the group to which it applies. In this way, we hope to make the specific group, time, and location to which a cultural observation applies as transparent as possible. When the same variable has been coded at multiple points in time, it may be possible to consider cultural change. To account for observer and coder bias, each datapoint is also tagged with a primary sources (e.g., source ethnography). We recommend users assess code reliability by returning to the primary source(s), and consulting additional sources, whenever feasible.
We have not attempted to combine cultural data across different datasets, for two reasons. In the case of cultural variables, different coders/authors often used slightly different codes, coding scales and/or coding rules. In the case of societies, different authors often coded data for different time and place foci for the “same” society. Because cultural practices change over time and vary by region, code discrepancies are to be expected when these foci are different. For further discussion of these issues, please see the section “Combining cultural data across the EA and Binford datasets” in Kirby et al. (2016). Despite these caveats, we have attempted to make it as easy as possible for users to identify similar variables and closely related societies across datasets. In the case of variables, users can search the variables table by keyword or theme; after selecting variables and considering the coding rules, decisions on when and how to combine data from the different datasets can be made. In the case of societies, we have assigned societies that share a focal location (though not necessarily a focal time) a shared cross-dataset id (xd_id). In addition, where justified, users might decide to group societies based on shared dialect and/or language. The societies table allows societies to be filtered by dialect/language.
We would love to hear from you. Please feel free to contact us directly, or to post a comment to our public GitHub issues page. D-PLACE aims to make already-published information on culture more accessible – previously, many of the datasets in D-PLACE were available only in difficult-to-access academic journals, or as downloadable spreadsheets on the webpages of researchers working with the data. In publishing the data again here, we have tried to make as many of the ‘details’ surrounding each datapoint accessible with the data. For example, each record of a cultural feature is tagged with the specific time, place and subgroup of people to which the record applies. We have also tried to make primary sources easily available to D-PLACE users, so that observer and coder bias can be considered together with a given cultural ‘code’.
We welcome contributions to D-PLACE, including both corrections to data and new contributions. Five types of contribution are possible:
Please contact us if you’d like to discuss a contribution, or would like to link your dataset to D-PLACE. You can also add suggestions for cross-cultural datasets that should be included in D-PLACE on GitHub – just open a new “Issue”, and tag that issue with the “Dataset” label.
The language spoken by a society is an important indicator of historical relatedness, cultural identity and contact. D-PLACE specifies the broad language family affiliation for all societies using the classification systems of Glottolog. Users can treat language family as a variable of interest itself, or can use it as a coarse-level control for relatedness among societies.
D-PLACE is carefully documenting data corrections, of which we have had to make many. In compiling some of the ‘core’ D-PLACE datasets, we decided to “undo” decisions by early dataset digitizers to aggregate or ignore some codes, and to simplify variable and code definitions. (These early digitizers were often working with punch cards, so we understand the desire to simplify!) In the case of the Ethnographic Atlas, we re-digitized the original data tables from Murdock’s 20+ installments, each published as an article in the journal Ethnology. Many users have written to us to provide corrected geographic coordinates for particular societies (and we continue to work on improving the accuracy of lat/long coordinates for some datasets). We also have a number of linguists on our team, and have worked carefully to verify that society-language matches are correct. For more on the approach we have taken to digitization/dataset correction, please see the supplementary information of Kirby et al. 2016. More recent changes are being documented on our GitHub site, and therefore are publicly traceable. Finally, D-PLACE is being versioned and releases archived and accessible through ZENODO, so any analyses that rely on D-PLACE can cite the particular version of the data that were used, enhancing replicability.
The folders in our data repository on GitHub include raw CSV files that can be downloaded and manipulated outside of D-PLACE. Please note the version of D-PLACE you download, as we are constantly correcting the datasets.
If you work with the Python programming language, you may also use the Python package pydplace for programmatic access to D-PLACE data.