Coin Hoard Analysis

It has been nearly a year since any update whatsoever on the development of Numishare to disseminate and visualize information about coin hoards.  In May 2012, just before the Nomisma.org/European Coin Find Network meeting in Frankfurt, I briefly discussed the new XML schema I had been developing.

That post described our use of the highcharts javascript library to render typological information in the form of graphs.  That blog post contained an image of a column chart with the total occurrences of coins per issuer in a single hoard.  This basic functionality has been expanded in several ways:

1. You can now compare up to six hoards at once with the chart visualization
2. You can download data as a CSV, with no limit to the number of hoards one can select for this comparison
3. Percentage of total has been added as a numeric response type (as a default) in addition to raw numeric count per typological attribute. This is a more accurate gauge of a hoard's contents, that is to say, 10 out of 10 coins (100%) being aurei is more telling than say, 12 coins in a hoard of 12,000.
4. Visualization based on date has been introduced, with an option to view as a cumulative percentage.
5. Charts can be printed and exported as JPG, PNG, PDF, and SVG files.

These features are now available in the trunk of the GitHub repository, and will be available to use when Kris Lockyear's Coin Hoards of the Roman Republic Numishare-based project is launched within the next few weeks.  This launch will take place before Kris and I head off to Perth for CAA 2013, where we will be presenting on the project: its evolution from dBASE III to XML/linked open data utilizing concepts defined by nomisma.org.

These analyses have enormous potential, and we are still at the tip of the iceberg for what is possible to do with these data. In the very least, these features enable scholars to conduct the quantitative analyses they have been accustomed to do for decades, only they may be executed in seconds instead of hours or days.

The thought processes that have gone into developing the XML schema and the development of Numishare to handle hoard data will be detailed in my MA thesis, which I plan to release openly and freely when I graduate in May.

OCRE en français

This week I have dedicated time to making Numishare more flexible with respect to switching interface languages.  Numishare has been able to use a "lang" URL parameter in record pages to display labels for Nomisma-defined numismatic concepts in alternative languages  for quite some time (by matching the lang param with the xml:lang in the skos:prefLabel in Nomisma's RDF).  However, most labels in the HTML interface were hard-coded into the XSLT.  Labels are now resolved through an XSL function which replies with the correct textual strong based on the field name and "lang" parameter.  While the Solr facets are still in English, the application is well on its way toward greater fluidity in transition between languages.  The next version of OCRE, a branch in the new Numishare GitHub repository, will support French, and possibly Arabic, thanks to the work being done by the ANS's collaboration with the Egyptian National Library to deliver its collection of Islamic coinage.

This work would not be possible if not for Nomisma.  When Nomisma supports an increased array of labels in alternative languages, it will be possible to provide Numishare public interfaces to a wider international audience.

The next version of OCRE will be available this spring, and will likely feature more physical specimens and findspots.

edit: By the way, the test version of OCRE pulls data about physical coins and hoards from a Fuseki SPARQL endpoint, an advancement that deserves its own dedicated post soon. The above screenshots show a widget in action that reads SPARQL query results and renders them in OCRE record pages and search results.

Numishare has moved to GitHub

After nearly four years of development and 400 subversion commits on Google Code, Numishare has moved its code base to GitHub.  The migration has been a long time coming, as nearly every other ANS project (except EADitor, which, like Numishare, also began during my work at the Scholars' Lab of the University of Virginia Library).  The move to GitHub will allow Numishare to be maintained more effectively in the long run, as its branching and merging mechanisms are much easier.  At the time of these blog post, there are 7 or 8 Numishare projects: some already reflected with branches within the main Numishare trunk, some with their own repositories.  Moreover, Numishare has a much more effective ticketing system, making it easier for me to address bugs and feature requests.

The URL for the repository is https://github.com/ewg118/numishare.

Improved Visualization: Comparing Queries

Recently, I began working on improved analytical mechanisms for coin hoards in Numishare.  These have not yet been completed (and so I will post about them later), but I put hoard work on hold for a few days to implement improved analyses for coin and coin type collections.  The updates can be seen in the Visualize page in OCRE.

The new features allow for selection of different graph types, numeric results in percentage of total or total count, custom queries (in addition to facets), and comparisons with other queries.

These visualization parameters are stored in URL parameters RESTfully, so it is possible to bookmark and share the charts that you generate with others.

Documentation: http://wiki.numismatics.org/numishare:visualize

Examples

Percentage of a selection of denominations issued by the emperors ranging from Augustus to Antoninus Pius in OCRE. Hadrian issued nearly 60% of all cistophori; Claudius issued about 40% of all didrachms and Nero issued nearly half (http://bit.ly/11xQMvH:

Comparison of Augustan coins which reference the Parthian settlement with Augustan references to Actium (http://bit.ly/TLJPRh):

New APIs introduced to Numishare

Over the last week or two, I have devoted some time to developing a systematic (and documented) set of APIs for querying and extracting data from Numishare.  These advances have been applied to OCRE already and will be available in the new version of Mantis, due out in the near future.  Moreover, since I have heard one or two requests about actually downloading data en masse from Mantis, I developed a rather simple PHP script which allows the user to do this.  The PHP script has been added to the Numishare codebase in tools/get-ids.php.  The comments of the script contain the basic execution instructions.

Documentation for Search API: http://wiki.numismatics.org/numishare:search

Documentation for Get API: http://wiki.numismatics.org/numishare:get

The Search API documentation details the two response formats (Atom and RSS) and the Lucene query Syntax (since the search results are derived from Solr).  Most importantly, the wiki lists all of the fields which are used for searching and sorting and their data types, as well as several usage examples.

Antoninus Pius added to OCRE

Today, the American Numismatic Society is pleased to announce that all coin types of Antoninus Pius from Roman Imperial Coinage (RIC) have been added to Online Coins of the Roman Empire (OCRE).  This increases the total by 2,054 to 10,536 imperial coin types from 27 B.C. to A.D. 161.

OCRE – A major new tool for Roman numismatics

Today, in collaboration with New York University’s Institute for the Study of the Ancient World, the ANS launches a major new tool to aid in the identification, research and cataloging of the coins of the ancient world.

OCRE (Online Coins ofthe Roman Empire) is an attempt to present, in an easily searchable form, all the varieties of the coinage issued by the emperors of ancient Rome. Phase 1, which is launched today, covers the coinage of the first emperors, from Augustus to Hadrian (27 BC – AD 138).

The site presents a basic description of each published variety based on the ANS’ collection catalogue (MANTIS). Each of these type descriptions is linked to specimens present in the ANS collection and, where available, to images. Searches are made straightforward through a series of facets, presented in a way that will already be familiar to users of other ANS search tools.

Traditional searches of familiar numismatic categories such as obverse and reverse legends and types are provided, in the hope that OCRE will provide an identification tool useful to collectors, dealers, curators and field archaeologists.

Subject searches have also been provided to allow more general researchers to find personifications, deities and portraits.

“OCRE is yet another example of the way that the ANS is both presenting numismatic material to those knowledgeable in the field, as well as expanding the accessibility of numismatic material to broader audiences”, notes ANS Director Ute Wartenberg Kagan. “Building on years of curatorial work to catalogue our coins, we hope that our new web-based tools will make that work available to as broad an audience as possible, in as flexible a way as possible”.

ANS database developer Ethan Gruber, who built OCRE, explains how it has been designed from the beginning to use a Linked Data approach to deliver added functionality: “OCRE is built on Numishare, an open source suite of applications for managing and publishing numismatic collections on the web.  The underlying data model of the collection is the Numismatic Description Standard (NUDS), a linked data-influenced XML ontology for coins.  NUDS enables the linking of coin types in OCRE to numismatic concepts represented on Nomisma.org as well as linking to web resources that describe physical specimens, such as those in the ANS' own collection.  Data about these specimens –images, weights, findspots–can be extracted for statistical and geographic analyses in OCRE.” A key element in the design has also been to link other stable resources describing the ancient world, such as Pleiades project for ancient geography.

OCRE project manager and Roman specialist, Gilles Bransbourg describes the advance that is heralded by OCRE: “OCRE is a leap forward for numismatists, historians and archaeologists alike. Until now, any research into Roman imperial coinage had to rely on paper-based catalogues, online auctions or the very few collections available online. OCRE offers a single, central online catalogue that allows users to view, download and organize digitized information covering the entire history of the Roman imperial coinage. The attraction of OCRE is that it is built as an open system. Any significant public or private collection may now link to OCRE and make its coins available to the wider public. Coin types will be connected to a growing number of examples from an ever-expanding number of sources. The digitized availability of relevant information like weights, modules, materials, legends, images, issuers, mints, location of find, and finally pictures, opens vast fields of research in many different directions and will hopefully inspire other areas in numismatics and beyond.”

ADDRESSES

OCRE: http://numismatics.org/ocre

Other ANS research tools:

MANTIS (the collection database): http://numismatics.org/search

ARCHER (the archives database): http://numismatics.org/archives/

DONUM (the library catalogue): http://donum.numismatics.org/