Visualization goals
Migrant farmworkers and allies have been organizing to expose conditions under Canada’s seasonal migrant worker program and to win immigration status and legal protections. With farms scattered across rural areas, it is challenging to reach workers. While organizers have built up geographic knowledge through years of outreach, I wanted to create a map drawn from an official dataset of all farm locations.
My goal was to build an exploratory map that could support organizers to see concentrations of migrant workers working in specific regions and for specific employers in Ontario, and in doing so identify new patterns and possibilities for organizing.
Dataset
The Government of Canada provides a dataset of employers with approved migrant worker positions (LMIAs) that includes name of employer, city, postal code, program type, industry code and number of positions granted. I cleaned up the dataset in Open Refine by separating the city, province and postal code into 3 columns and eliminating duplicates.
Other examples
I found a few other maps created using this dataset, although with a very different objective. The goal of these maps seems to be to stoke zenophobic narratives about employer reliance on migrant labor. The existing maps also have design and data limitations. The first map, while visually simple, only shows the number of approved LMIAs. On the second map, it is difficult to see the number of workers or employer name and overall patterns are hard to distinguish given an awkward zoom function and unclear scale of the circles. Neither map is designed for organizers to guide outreach strategies.
First iteration
To show concentrations of farmworkers by region, I created a map with filters for the agricultural program and specific industry codes. I filtered for employer names in a second map. For both maps, circles were layered on top of each other, making it difficult to see the full picture of the data when there were multiple circles in the same geographic region.
My initial fix was to create filters within the first map (and an accompanying chart of top regions) to be able to isolate specific areas in the second map, and to add a drop-down menu where the specific employer names would show up. However, occlusion was still an issue for the map visualization itself.
I feel I did accomplish in part my goal of visualizing data to illustrate broad concentrations of migrant work, and the ability to drill down to specific employer names. I could immediately see some initial patterns. For example, 4 of the 5 top cities were expected based on field knowledge of major areas of migrant work. However, the city of Schomberg was unknown to me, and it was notable to see just one employer accounting for a sizable concentration of work permits. As an organizer, I would want to understand if there has been contact with workers at this employer, and if not, to prioritize outreach there.
User feedback
I asked 3 users for feedback on my initial dashboard including whether the map was legible, what would make it more user-friendly, whether they could see using the map to support any current work, and what would make it more useful for their organizing.
I reached out to two people who are organizers working with migrant workers, and one non-organizer. The feedback from the first user (an organizer) was that the map was very easy to understand, but that the maps could be bigger and that the dots were hard to distinguish. She said that the dashboard could help with research on farms and legal cases, and suggested layering in more data such as worker nationalities and the crop type. The second organizer’s feedback was simply “Wow this is great.”
The third non-organizer flagged that they didn’t understand the terminology and the dashboard needed more context. They asked for more guidance on how to move through the dashboard and as it was unclear how the two maps were connected. They suggested making the trends about the biggest employers clearer and eliminating the industry code. They also suggested making the geography easier to decipher, and the employer drop down list bigger.
Second iteration
To integrate this user feedback, I created larger maps in my dashboard and used a “scrolly-telling” format. I made the maps the width of the dashboard, and moved some related elements (like a chart with the top cities with migrant workers) into the unused space in the map.
I added more context in the introduction to make the maps more accessible to a broader secondary audience and larger header questions to guide the user. I also provided more specific instructions on how to interact with the maps.
I removed the industry code filter and I added a chart with the top employers as part of the context to introduce the second map. Finally, I changed the map format to make the water more distinct, and enlarged the employer drop down menu.
Next steps
I was glad to hear the initial positive feedback from organizers and I think the iteration through the user feedback process made the maps easier to navigate, and made the overall design more visually interesting.
However, the occlusion issue still makes this map difficult to use. I would like to try the stacked bubble suggestion offered by Professor Sula to see if that could improve the ability to see the overlapping circles and overall patterns, and find employer names. As an immediate fix, I increased the size of the circles and the opacity, and gave them a light grey border (as shown in the updated maps below). I also tried mapping locations to city instead of postal code in the first map. This does reduce the number of overlapping circles in a given city or town, but makes the locations less specific.
I would love to speak with organizers in more depth on how to further build out the dashboard to meet their needs. My hope is that building an initial map can create a template to collect and layer in more data that doesn’t yet exist in a dataset. I would love to co-design this template with organizers and workers to link their knowledge drawn from organizing in the field with this list of employers.
A future design could layer in other available data (such as labor violations), and provide more information for each employer in the tool-tip., I would also love to include case studies of worker organizing at specific employers as part of the overall dashboard narrative and context.