‘Cultural genocide’: the shameful history of Canada’s residential schools – mapped

Country/area: United Kingdom

Organisation: The Guardian

Organisation size: Big

Publication date: 06/09/2021

Credit: Antonio Voce, Leyland Cecco, Chris Michael


Antonio Voce is an Interactive Journalist for the Guardian.

Leyland Cecco is a reporter for the Guardian based in Toronto, Canada.

Chris Michael is an international news and series editor for the Guardian, and host of the Reverberate podcast.

Project description:

The aim of this project was to instruct readers (especially outside of Canada) on the historic context which led to the recent discoveries of several unmarked graves at the sites of former Indigenous Residential schools in Canada.We wanted to make sure readers got an understanding of who the indigenous people of Canada are, before telling the story of forced assimilation by the Canadian government.

We also wanted to have an “evergreen” dashboard that would be updated with the latest discoveries, and that readers would be able to look up again without scrolling throught the whole story again.

Impact reached:

The story was widely shared in and out of Canada and ended up in the GIJN’s top 10 data jouralism projects of 2021 list.

Techniques/technologies used:

 This is a web project so the core technologies are html and javascript.
– we created a a hillshade map of Canada by cutting and editing a raster layer in Qgis.
– we extracted the cultural area boundaries shapefiles from native-land.ca, and rendered them using html canvas. All interactivity is done in javascript.
– we obtained the country’s boundaries and historic treaties shapefile from the official Canadian open data portal.

– we obtained the dataset on residential schools from the National Centre for Truth and Reconciliation and added an svg layer to render them so that they would be easily hoverable.

– After talking to survivors about the long journeys they had to endure to reach the schools, we recreated them in Google My Maps, then exported them as geojson, rendered them and animated them with html canvas.
The choice to use canvas, which is trickier to work with, rather than easier tools like d3, was made to ensure that the visuals would run as smoothly as possible.

– Archival photos from the schools were selected and cut in Adobe Photoshop.

– FInally, we used a in-house scrollytelling library to let readers scroll through the story and trigger the various steps of the visualisation.

What was the hardest part of this project?

The hardest part was to create empathy in the reader by keeping focus on the people who are the main subject of this story. It is easy to loose track of this aim on a data-led story such as this one. You want to keep the attention focues on humans when all you have is dots on a map and some numbers. We tried to achieve this empathy by choosing vibrant colours for the different cultural area shapes, by using archival photos portraying indigenous children, and by animating school-home distances to make them look more like journeys.

We also tapered the story so that the “macro” would show at the beginning (the map with all the findings) and then progressively shifting the focus towards the individual, with Jack Kruger’s memories, the children’s photos and the zoom on the St. Eugene school.

When working on data stories, and especially on a dark story like this one, it’s important to always remember you are talking about people, not numbers or data points. We tried to dignify the people through the data, the choice of palette and tone. We wanted people to think of the victims and to learn and think more about Canada’s dark past.

What can others learn from this project?

Hopefully that visual stories aren’t and shouldn’t be for data nerds only, and that cool visuals are only a means, not an end to a story. It doesn’t have to be flashy or very complex as long as it conveys the story you are trying to tell. And engagement and attention time will always be higher if the story sounds like it’s talking about people and human experience, rather than “data”.

Project links: