The aim of the "Laboratory Prototype" project was to develop a "proof of concept" which would demonstrate the viability and utility of the concept at a technical level. The general goal of the prototype was a publicly-accessible web service – a kind of Facebook for dead people – with distinct community and household-level applications, and the first phase was to develop the engine.
We began by developing household-level property data for two historical communities, the "Old North End" of Halifax, and "Durham & the West River of Pictou", with the initial time set at 1881. These communities were considerably different in character then, as they continue to be today: the Old North End of Halifax was already urban in character at that time, and residents worked largely in the trades and small business. The area was composed of a core population of Irish Catholics, and contained many households with unrelated members (roomers, co-tenants, recent immigrants, etc.). Durham, the intellectual and market centre of the West River region of Pictou County, on the other hand, was rural in character, with agricultural occupations dominating. The Village was a centre of Scottish Presbyterianism, and was composed largely of family households, often extending across several generations.
The MIRCS Institute conducted the work on the Old North End of Halifax and the Durham Heritage Society pioneered work in the West River district of Pictou. This work yielded several hundred households in Durham and about four thousand households in the Old North End. Using those names, in combination with other local knowledge, we were able to connect those properties with the 1881 census records, almost in whole for Durham, and with a partial sample for the Old North End.
As the historical data is organized by household and by community, we have the opportunity to develop community aggregates and household profiles at each geographic level. Our expectation for the full-scale platform is that the community aggregates would be freely available to the user, while the household profiles would require a user fee. Interactive views of a community would be provided on a searchable map, from which navigation could proceed to household profiles for specific detail. Household profiles would be derived based on joining the different data components, and the GIS capabilities provide the means to compare immediate neighbours, physical distances to other households with similar attributes, and the like.
We believe that the key to releasing community energy and action for genealogical work is to minimize the data framing required, and allow community efforts to focus on the collection of the data itself. In other words, the work of creating data structures and appropriate frames should be understood as programming problems to be solved independently. The matter of incorporating multiple datasets constitutes the core of the programming problem. The chief programming work has sought to resolve the creation of dynamic databases for the full-scale platform. Community-based historical and genealogical groups in the province have built extensive archival collections in recent decades, with sometimes quite extensive digital records, but these digital datasets use different software bases, data structures, source authentication, and quality standards. Given the volume of digital information available, we conceptualize it as a programming problem: how to join arbitrary data from disparate research groups with weak data management expertise of their own.
We conducted a series of projects with five course-based, infomatics and computer-science student teams from Dalhousie University. In conducting these, we have tried to solve for the database programming and the mapping API at the same time: the solution needed to incorporate multiple datasets, and provide interactive mapping capabilities. Ideally an end user would be able to easily navigate a historical neighbourhood in an engaging way, and drill down on specific households for a detailed profile. The first three teams helped in the work to provide a searchable, easy to use, Historical GIS application and a report on progress was presented at the 2017 ACM-SIGDOC conference (Johansen, Sundararajan & Armstrong, 2017). Since then, we have conducted further work with two student programming teams which has brought the "build" of the core programming engine to completion.
This is a large and ongoing project, and as we have outlined, we have completed the technical "build" of a laboratory prototype. The next phase of the project involves advancing the data visualization and mapping problems, and integrating them with core programming. We call this phase "Field Trials".