Montréal, 1900-1905. Rue Craig >Est, depuis rue McGill, BANQ Carte Postale, Photo: Flickr

New ways to imagine an old city

Montréal, 1900-1905. Rue Craig >Est, depuis rue McGill, BANQ Carte Postale, Photo: Flickr

Imagine if you could examine the entire real estate portfolio for any proprietor in a large city linked to detailed household descriptions of up to a third of his or her tenants. Imagine if when you did, you discovered that women owned a quarter of all rental units in the largest city in turn-of-the-century Canada and that they appear to have managed these properties differently. Wouldn’t that change how you think about gender relations in the past?

This is but one example of the remarkable potential for novel imaginings arising from the latest phase of the research infrastructure Montréal, l’avenir du passé (MAP), Canada’s oldest and largest historical GIS.

This exceptionally rich resource consists of four distinct elements: a new cartography of all properties in the city in 1903, detailing who owned what; an index of all household heads in the 1901 census linked to this map at the lot level; a 30% sample of the complete manuscript census returns of the city’s households; and a geo-referenced vector map of all 101,353 buildings in the city in 1912.

These research tools for understanding Edwardian Montréal build on MAP’s earlier layers for 1880, 1846 and 1825, which are available online. The combination permits detailed study of both the evolution in land use and population densities over a century. The transformation of our relationship to the rest of nature is strikingly evident. So too is the development of qualitatively new suburban landscapes, where the environment that mattered most was one’s own family.

The map for 1903 is based on a 1,360-page publication by the City Council listing all properties in the city, which as the preface put it was “designed to act as an incentive to the sales of property.” For this to work required proper identification of each of the 11,322 individual proprietors, as well as the 469 corporate or institutional owners. In addition the square footage, approximate dimensions, rental value and assessments of buildings and land for each property were given, as well as the civic numbers for built properties.

An innovative aspect of this publication was its use of highly revealing standardized pricing for evaluating land, ranging from 2¢ to $18 a square foot. Understandably the highest values were along Greater St James Street, the financial capital of Canada, but one can also see the emergence of the two, linguistically separate, bourgeois shopping precincts on St Catherine Street. On a more subtle level, the values on key cross streets and on corner lots speak to the significance of locally controlled economic relations in each neighbourhood.

Land values per square foot, Montréal, 1903. Map Credit: Robert C.H. Sweeny and the MAP team.

Land values per square foot, Montréal, 1903. Map Credit: Robert C.H. Sweeny and the MAP team.

Using the descriptions for each of the 32,152 entries on this roll, in conjunction with the 1897 Goad and the 1907 Pinsonneault atlases, I mapped 30,799 individual properties. It was not unusual for a single entry to represent multiple adjacent lots, while frequently complex industrial properties consisted of numerous entries on the roll but had much of the data listed under a single entry. This was most evident with the railway lines.

To people this map we linked the index of household heads to their respective lots. We are still in the final stages of verifying this work, but we have already linked 98.4% of all households for whom the census provided a civic address. This remarkable level of linkage was possible because we base our approach primarily on location, rather than names. Each census enumerator developed their own spatial approach to their task, which is reflected in the order of households in the manuscript returns. This internal spatial ordering, analysed in conjunction with street and alphabetical listings of the 1901 city directory and the street numbers in the Pinsonneault atlas, has permitted us to raise the initial highly respectable linkage rate of 84% based on geo-coding to almost complete coverage.

This index provides information on 70,076 households on the island of Montréal, 51,758 were within the 1901 city limits. Compiled by genealogists through a collective, volunteer effort, this index provides varying levels of information. For the city itself and in descending order of coverage, it contains data on: the size of 99% of households; the name and surname of 99% of household heads; the civic address of 96% of households; the number of rooms in 90% of the dwellings; the employment status of 32% of household heads; the income of 25% of households; the presence and number of servants in 9% of households. Used with caution, this admittedly limited descriptive information can be very helpful in the early stages of developing a research project, because all of this data can now be viewed spatially.

In 2007, Sherry Olson asked colleagues whose research projects had involved the 1901 census for Montréal if they would be willing to contribute their databases to a combined sample. Five projects accepted her invitation: Peter Baskerville and Eric Sager provided two, (the Canadian Family Project as well their earlier sample for the study of the unemployed) Danielle Gauvreau and Peter Gossage’s demography project; Mary MacKinnon’s project on wages and employment; and Patricia Thornton’s collaboration with Sherry on infant mortality. The differing objectives of these projects combined with their varying methodologies meant there was a considerable amount of work in editing and completing data entry so that the file contained all the returns for all of the categories in the sampled households. The result was, however, more than worth the effort as 30.7% of all households in the index are included in this combined sample.

Percentage of Households in the MAP 1901 sample. Map Credit: Robert C.H. Sweeny and the MAP team.

Percentage of Households in the MAP 1901 sample. Map Credit: Robert C.H. Sweeny and the MAP team.

This is not a random sample. As can be seen in the spatial coverage by ward, the newer suburban wards beyond the limits of the city in 1880 are distinctly under-represented, while the famed “city below the hill” of St Anne ward received the greatest attention. Furthermore, each project used a differing sampling technique, so one has to use care in selecting which records to include for a particular query. Nonetheless, the sheer size of this new dataset means we can answer more detailed questions, and, thanks to the spatial linkage and the possibility of using the index, doing so in a far more contextualized manner. For example, using this combined sample I was recently able to compare the housing patterns of the newly arrived Sephardic, Cantonese, Syrian, Newfoundland and Italian communities. This simply would not have been possible working with what was previously the largest (5%) publicly available dataset.

The final element of MAP’s new release is our most accurate geo-referencing of the built environment of Montréal. Aided by several students and with technical advice from Don Lafreniere, over the past three years Sherry Olson created this 21st century view of the 1912 Goad atlas. In our previous work, we drew polygons for each of three layers: buildings, lots, and blocks. This led to numerous inaccuracies, as buildings did not necessarily line up properly with lot lines or other features we had drawn. For 1912, we used a different technique. We simply drew lines, which in GIS are called arcs, everywhere there was a line on the original map. We had the GIS software combine these lines into all of the possible polygons and then we selected and combined the thousands of resulting shapes into separate layers for buildings, lots, blocks, etc. This technique was faster and more accurate as features that shared a line on the original map are constructed from the same arcs, eliminating any possible overlaps or slivers.

The built environment in Montréal, 1912. Map Credit: Sherry Olson and the MAP team.

The built environment in Montréal, 1912. Map Credit: Sherry Olson and the MAP team.

We have yet to people this map, but in its rich detailing of the 49,967 buildings covered in brick or stone and the 50,458 in wood, there is a great deal that one can do with it already. The dramatic transformations wrought by the railroads presage the gradual eclipse of the century-old horse-drawn economy. While the contradictory processes of much higher population densities and significantly increased space for working class, as well as middle class, families speak to the social and cultural choices creating this new urban environment.

All four elements of this most recent addition to MAP are currently in beta-testing. We plan for full release of the finished products at the Canadian Historical Association meeting in Toronto in May. Graduate students who cannot wait that long and are willing to work with the beta versions should contact me directly.

Print Friendly
The following two tabs change content below.
Robert Sweeny is a socio-economic historian, who lives in St John's Newfoundland with Elizabeth-Anne Malischewski. His monograph Why Did We Choose to Industrialize? won the 2016 Governor General's award for scholarly research. He is currently working on gender, property and national identities. A Professor of History at Memorial, Robert has been involved in the H-GIS project Montréal, l'avenir du passé since its inception in 2000.

Latest posts by Robert C.H. Sweeny (see all)


  1. Anthony Chadwick says:

    Such a wonderful resource! Should keep grad students and professional historians busy for years to come!

Leave a Reply