Thematic Mapping

The data for today’s lesson is from the public history project Montréal, l’avenir du passé (MAP), directed by Sherry Olson and Robert C.H. Sweeny. Translating to Montréal, The Future of the Past, MAP digitizes, geocodes, and georeferences data from historical atlases, city directories, census records, and municipal tax rolls to show demographic change over time and space in Montréal. For more about the MAP project

As a public history project, MAP makes their spatial datasets freely available and their entire website and documentation is bilingual (French & English). Not only that, they have prepared QGIS applications which visualize selected data. You can simply download one of these “apps” and explore a given time period. MAP is therefore an exceptional example of digital humanities using GIS! And QGIS, for that matter!!

MAP has enthusiastically encouraged us to use their data in this DHSI workshop, but we highly recommend exploring their datasets and prepared QGIS applications directly. Today’s exercises will work with data available in the Goad 1880 applications which “build on a rectified edition of Charles E. Goad & Co.’s atlas of 1880 created by MAP.” This download contains multiple QGIS projects and data all from 1880-1881. While physical features such as water, streets, and buildings come from the 1880 Goad Atlas, administrative data on lots, wards, and census tracts — and of course census data itself — come from the 1881 Census. Data on occupants and business rents comes from the 1880 Municipal tax roll, and further demographic information comes from Lovell’s directory of 1880-81.

More about the MAP project can be found on the Centre interuniversitaire d’études québécoises (CIEQ) website: https://map.cieq.ca/. This site is tailored less towards downloading and playing with datasets than exploring them online. For instance, mix and match variables, how people made a living, street crawling, or just look at Lovell Directory data in web map form!! For more on their sources, see the CIEQ website subpages “Our Gorgeous Sources” and “In the Storeroom”.

What we’ll use…

Below we’ve outlined the data used in today’s lesson with a description of what it visualizes as well as, for reference, what it’s named in the Goad 1880 MAP application download and any modifications made to the original dataset. Data was renamed for ease of understanding.

All the data we’ll use is stored in shapefile format. While QGIS projects within the Goad 1880 MAP application download were set to the projection NAD83(CSRS) / MTM zone 8, individual shapefiles were not stored with projections but rather “projected on the fly” when loaded into a project. Therefore, all renamed shapefiles were also saved with the Coordinate Reference System (CRS) NAD83(CSRS) / MTM zone 8. The only difference this makes is in the kinds of sidecar files associated to a shapefile.

Demographic data from 3 sources

census1881 Demographic data from the 1881 Census containing 111,208 features, with information on name, age, sex, occupation, religion, disability (‘blind’, ‘deaf’, ‘dumb’, ‘unsound’), birth place, and more. The geographical tie is at the lot level, and while the full census contained some twenty thousand more responses, MAP was only able to successfully link 111,208 at the lot level (quite impressive!). This data layer is named g80cen81 in MAP’s Goad 1880 application download.
lovell_directory Demographic data from Lovell’s Directory, 1880-1881. This directory contains 34,658 features with attribute information on name, sex, occupation, address, widow status, and kind of living quarters. There is also information on whether each entry was part of a firm/company, and whether they had a second address. As MAP writes: “Although many of these people gave both their work and home addresses, most entries are for where people lived.” This data layer is named lovell_mtl in MAP’s Goad 1880 application download. This is not to be confused with the layer lovellmtl which contains practically the same information but which doesn’t break down occupation into distinct column(s). Lovell’s Directory in web map form can be explored here.
taxroll_occupants Demographic data from the Tax Roll of June 1880, City of Montreal. This dataset contains 29,453 features with information on occupants’ first and last names, address, sex, occupation, rent, and ward. Tax Roll data is limited to heads of household meaning there are fewer entries than Lovell’s Directory and far fewer than the 1881 Census. MAP sourced the data from Archives de la Ville de Montréal. This data layer is named occupant in MAP’s Goad 1880 application download. Note however that the Goad 1880 application download also contains a file called g80_roll which contains the same number of features. Indeed, it is almost identical (though it contains more attributes). However, there are a few inconsistencies between them. Despite this, occupant was chosen over g80_roll as it has fewer attributes and for the purposes of our QGIS practice, minor discrepancies are of no concern.

business_rent Rent paid by businesses according to the Municipal Tax Roll of June 1880. This data layer is named bus_rent in MAP’s Goad 1880 applications download. It contains only 5,117 features, whereas MAP’s dataset on residential rents (res_rent) contains 25,457. Combined, this amounts to 30,574 features which is more than the Tax Roll dataset of 29,453. This is simply because business proprietors who were also heads of household would have submitted entries both for their place of residence and their place of work.

Administrative Boundaries/Units

lots Individual lots according to the 1881 Census. Containing 12,226 features, this layer was originally named censuslt in MAP’s Goad 1881 applications download. censuslt is geographically nearly identical to newlots81, the latter containing one more feature. However, the two datasets differ in attribute tables. censuslt was chosen for our exercises because it has a fewer attributes and thus a cleaner attribute table. Both censuslt and newlots81 differ from the dataset for lots used in MAP QGIS applications which visualize g80_lots. This dataset has a few hundred fewer lots, and differences are especially apparent around the waterfront. Because we will be visualizing census data, lots for our workshop are based on census lots.
wards Wards from the 1881 Census. This data layer is also named wards in MAP’s Goad 1880 applications download.
census_tracts Census tracts 1881 Census. This data layer is named ctracts in MAP’s Goad 1880 applications download. To see visualizations specific to this administrative/census level see the QGIS Project Census returns 1881.qgs in MAP’s Goad 1880 applications download.

Contextual Features

Physical and infrastructural features are organized into the data subfolder features. These layers are parks, beaches, lanes, streets, quays, and water. Each shapefile was slightly renamed for consistency and saved them with a projection. Additionally, the water dataset was modified to rectify an issue with the geometry in the canal feature.

You’ll notice when you open the QGIS project for today that the files for contextual features are already loaded — if hidden — in the project. This is because MAP went to great efforts to curate a symbology for this basemap data that works well, and re-loading them to our project would require we do all this work over again. This way, we can simply turn on the layer group when we are ready to map.

Alright! Now that we understand what data we’re using and where it came from, Let’s move on to some hands on practice with QGIS!