Statistics Canada has released its Population and Dwelling Counts from the 2016 Census. This is the first of several releases scheduled for StatCan – it plans to release data products for all modules (e.g., income, labour, ethnocultural diversity) before the end of 2017. Hats off to them!
This is a map I quickly rendered in Tableau. I’ve played with Tableau in the past but usually stuck to hacking out code and data myself. After a December meeting and tutorial with Lucia Costanzo at the University of Guelpgh Library, though, I’ve become a bit of a convert. The proprietary nature of the software is sometimes problematic, but it certainly speeds up the development of your maps and visualizations. Perhaps more on that later.
This map shows the population change in Ontario’s Census Divisions (CD’s) from 2011 to 2016. The percentage change is based on one census cycle to the next, but the color gradation ends up comparing this percentage change from one region to another. Note well that this can cause issues in interpretation of your numbers.
Some interesting things to note or remember:
This is considering population change only. It doesn’t take into account the value itself. So, while Kenora, in the northwest, shows the most significant population change, it might still be unfair to draw a comparison against any of the regions in the golden horseshoe or GTA
This week’s map visualizes Canada’s unemployment rates for October 2014, which were announced last week:
I coded this map with trepidation since comparing unemployment rates across provinces isn’t always as important as considering one province’s current rate against its own historic numbers. For example, this map shows us that Ontario’s unemployment rate still lags behind Alberta’s. No surprise there. What the map cannot do, though, is show that Ontario’s unemployment rate for this month – 6.5% – has finally recovered since the Sept 2008 crash. The last time Ontario’s unemployment rate was this low was in October 2008. To best visualize the province’s unemployment trend back to pre-recession numbers, one should simply chart the data, or even just give the real numbers in tabular format. The best way to do this on the web is with charts.js, which seems to be some of the easiest coding I’ve ever seen. That will be my project for later this week.
This week, I’ve taken the same population density variable I used last week and plotted it for Brantford, Ontario. I’ll be speaking about open StatCan data to our journalism students in Brantford in a few days’ time, so it was only fair to plot the same variable for our students in this city, too.
One of my sidebar projects this fall has been to get back into mapping socio-economic data. This is something I used to do quite a bit four years ago (these maps have sadly succumbed to linkrot and plugin abandonment). Projecting numeric data onto maps is easier than most people think, and ever since I moved to a new city in 2013, I planned to pick up this skill again to learn a few things about my new town. And as a data librarian, I know where to find and work with census data, so it was easy to kickstart things into gear once more.
The interesting thing about this map isn’t so much its colorful polygons, (based on statistics anyone can download here) but the tools I used to build it. When I was creating maps in 2010, the average person who wanted to hack something out was limited largely to using Arc on his or her campus, or using the open source (and still maturing) variant, QGIS, or working with Google Maps. These days, QGIS is very mature and has a strong developer community, GMaps is still going strong, and users can use services such as Mapbox’s TileMill. The options to choose from are stronger, and there is an option that can meet your background, whatever it may be.
As an example, I’m linking over to Mita Williams’s recent work mapping population change in Windsor, Ontario, as well as making the case for electoral change in her hometown. Mita is a UX librarian and far more of a coder than I’ll ever be, so her recent work with maps shows a freer hand at hacking out java to make things go, while I use plugins within QGIS to automate some of the coding for me, which frees up my time to spend on analysis.
At the end of the day, our maps are projected with the same code and with data from the same datasets, so our endpoint is the same, but the tools we’ve chosen to use may be better suited to our own particular abilities. That is something I didn’t see in 2010 as much as I see today. And that change is a good thing. Getting these datasets into the hands of the masses, and then making them usable and understandable for everyone, is crucial to the precepts of openness – open access, open government, open data – that we espouse as librarians. One can have completely open access to data, but its value is lessened when it cannot be used or understood by all of society. Yes, open data is a crucial part of today’s citizen-to-citizen and citizen-to-government relationships, but the more tools people have to work with that data, the better.
Link Sharing: This page links to a newspaper article on the potential impact of offsite storage for items with low circulation numbers in a province-wide academic library consortium. Check it out and contribute to the discussion!
One of the things I’m constantly doing as a government documents librarian is giving lessons on Statistics Canada geographic areas. Census geographies can be downright confusing to the new user (and to sometimes to the seasoned expert!). The names are riddled with acronyms and jargon, and their relationships to other areas and spaces can be complicated. One legally incorporated township may be considered a census subdivision while another may be classified as only a census agglomeration. Another city may be classified as a census subdivision, and also be part of a census metropolitan area of a similar name, e.g., Toronto CSD and Toronto CMA. Or, a city may be classified as a census subdivision and exist not only in a CMA with a similar name, but also a census division (I’m looking at you, City of Waterloo CSD, Waterloo Region CD, and Kitchener-Cambridge-Waterloo CMA). And if you dare introduce census tracts the first time through, your short introduction to the “Russian dolls” nature of census geographies runs the risk of turning your lesson into an information dump about privacy and data validity when all that your first-year economics student wanted to know was why it’s so hard to get comparable income and migration numbers for Kitchener, Ontario, and The Pas in northern Manitoba.
Confusion abounds. One of the problems we encounter are the tools we use to explain these geographies, which should be easily understood but are often abstract – we may live in towns and cities, but we refer to them as census agglomerations or CMAs. What can you use to show how spaces relate to one another, or how certain concepts can be measured and expressed spatially? The answer is a map, of course. God lov’em, those maps. Maps help us express numbers – quantities, amounts, rations, proportions – with colours and shapes, and in the regions we live in and travel through each day. Face it, “big data” wouldn’t be as big as it is today if we didn’t have “big maps” to help use make sense of the numbers. However, StatCan’s digitized maps are large, layered PDFs that aren’t always user-friendly. The Standard Geographical Classification (SGC) PDFs are great reference items, but they aren’t very accessible. And this creates a learning gap for so many of our users.
To overcome this gap, I’m constantly pulling out the old SGC print maps, and I’m also cutting and pasting and hacking together magnified screenshots of the PDFs into my slide deck. Typically, if you need census help and you’ve found me in person, then there stands a good chance that I’m going to crack open the SGC and unfold a map somewhere in the office (I even keep the southern Ontario CD-CSD map posted to a wall). I started doing this last Spring after I moved to Waterloo and had to learn the region’s geography and confirm its census divisions, subdivisions, and CMAs for myself, and I realized this was a simple and effective tool that should be used more often, especially with new StatCan users.
Typically, I bring students to a nearby conference room and unfold the map on a large table. I find that being able to “walk around” the entire map and point to the places where the lines that signify the different geographies merge, separate, and then merge again, helps students understand some of the logic behind the regions (at least in terms of distance and population). They may not always be able to recall all the differences between a census division, subdivision and metropolitan area after a session, but they at least remember that there are differences, and these differences are important enough to affect their research.
The classroom is a different story, though. When working with only one person or a small group, there is a persuasive element at work that captures everyone’s attention. Carefully unfolding and presenting a map to a small group of people is like opening a box that holds a surprise. (Let’s call this surprise “knowledge” and we’ll call ourselves awesome for charming our audience so handily into learning something). But if we take that same map into the classroom or lecture hall, it risks becoming an awkward, cumbersome prop. It can become a distraction or even a failed means to demonstrate your expertise in such a short time to such a large group of people.
Maps that unfold to become wider and taller than you put the room’s attention onto your map-wrangling skills (however good or poor they might be) instead of on the knowledge you have share, so I avoid them. You’ve never caught me walking to a classroom with a print map, and I doubt many other librarians do that today.
Instead, I give the class what they want and what they expect, and that means I work that map into my PowerPoint deck. Any time I’m introducing StatCan resources and geographies to a class, I insert three images of the same PDF map, each one magnified more than the last. This helps people “zoom in” with their eyes and see the many relationships and regions that are defined in one place alone. The length of time I spend on these slides depends on the classroom’s needs: sometimes, I spend only a few moments on these slides, and other times, I’ll spend five or ten minutes. What matters is that after I’ve finished up and am headed back to the office, I know that the instructor can pass around a slide deck that always refers to all these different areas.
I know I’m not presenting anything new in this post: maps have long been a tremendous tool within government documents librarianship. Perhaps the takeaway lies more in information literacy than it does anywhere else. Is your digital resource, as presented to you, the best way to help the user understand the resource? You may want to turn to the print resource or manipulate the digital resource, as I do with StatCan maps, to improve learning and synthesis. It’s just one more tool (or two, in this case) in our IL toolbox.
A few years ago, I designed a few rudimentary Google maps of Halifax from StatCan data. This was before I really knew anything about stats and data (n.b. I still don’t think I know much more than “some things” about stats and data), licenses, and how to properly interpret them. One map that I created showed Halifax’s population change, tract by tract, from 2001 to 2006. I’m giving myself embarrassment cringes by linking to it, but all the same: view it here.
Of note: the suburbs clearly rule the roost when it comes to Halifax’s population changes from 2006 to 2011. The only tract on the peninsula showing a significant increase (i.e., over 11.9%) is Tract 2050019.00, in the middle of the peninsula. The increase in this tract is due, I’m certain, to the Gladstone redevelopment, the first major phases of which were completed – if memory serves me correct – in 2007 or 2008.
For what it’s worth, I’m not sure if I’m going to build a google map from 2011 census tract data. The work is time-consuming and there are other people in my field who have the expertise and software to do a much better job than I can. (And besides, my own hobby at the moment has more to do with plotting historic maps with Google Earth!) My work finding socio-economic data, making the odd remark here and there, and helping others make sense of it, is enough work – and fun – for one person. 🙂
Finally, here are a few outbound links to keep you interested:
Are you coming to #CLA2011 (or #CLA11) in Halifax, Nova Scotia? Then this Google Map may come in handy. I created a Google Map to help a few librarian-friends from across Canada decide on some things to do in Halifax and then decided to share it with the world. Enjoy, contribute, and share and share alike.
And since you’re coming to CLA 2011, make sure you visit and say Hi! during Saturday morning’s Technology Lightning Strikes! panel at 8:30 (Session G49). I’m going to be speaking with a bunch of excellent librarians (read: absolute tech superstars who know so much more than me!) about emerging technology trends and how to integrate them into your everyday work with little fuss and hardly any muss. I’ll post more details on this in a later post.
This week’s map shows us 2005 median incomes for married-couple families in Halifax, Nova Scotia. Don’t let that long topic get to you: although Statistics Canada can sometimes get a little difficult with their language, it’s not too hard to decipher:
2005 median income – This is not the average income for the tract but the income that separates the top half of reported incomes from the lower half of incomes in the area. This is a commonly used value when considering income because it prevents incredibly high and incredibly low incomes from affecting a stated average.
married-couple families – StatCan records income for different family types. There are lone-parent families, of which “female-lone parent” and “male-lone parent” are subsets. StatCan also lists dual-parent families (my term). In these are two distinct kinds: married-couple families and common-law families. However, Statistics Canada does not combine these values for us into one field as they do with lone-parent families, so we must consider them individually.
Two interesting patterns emerge on this map. The first pattern is the manner in which lower median incomes become prevalent as one moves west to east. The further into old Halifax County one drives, the lower the median income will be. Presumably, lower rural-based incomes and dual-parent families who hold only one reported income between them account for this. Note, however, that in rural western Halifax county, we nonetheless find higher incomes: the incomes over extreme western Halifax are nearly double the incomes in extreme eastern Halifax.
The second pattern is the high incomes to be found on Halifax Peninsula and along the Bedford Basin. These incomes should be expected, given the socio-economic patterns we see in these areas (e.g.: highly educated, fully employed households). What is of interest, though, is the proximity of Halifax’s highest median income to its lowest:
Highest income for married couples in Halifax:
Tract 2050005.00 (which I’ve called South End-Gorsebrook), lying on the peninsula’s shores: $194,622
These tracts, nearly side-by-side one another on the Halifax Peninsula, house two distinct populations that are tied at the hip – the student underclass studying and working at the post-secondary schools and hospitals that dot the south end, and the professional class that is employed by these institutions. I’m painting with broad strokes here, of course, but it does serve as a little bit of context to explain how these two different income levels lie within only two or three kilometres of one another.
Today’s map is a Valentine’s Day treat for all the single ladies and men in Halifax, Nova Scotia. By manipulating 2006 Census data at the tract level, I’ve plotted maps that show the marital status of all the men and women in Halifax.(*)
1. Women who are not in a married relationship in Halifax, Nova Scotia, 2006 Census:
2. Men who are not in a married relationship in Halifax, Nova Scotia, 2006 Census:
(*) Careful attention must be given to meaning of these values. These maps represent the marital status of all people living in a tract, over the age of 15 – a question that was asked on the 2006 Census. When a person was asked this question, they could respond by stating that they were:
Never legally married (single)
Legally married (and not separated)
Legally married (but separated)
For the purposes of these maps, I have considered anyone who answered “Never legally married (single)”, “Legally married (but separated)”, “Divorced”, or “Widowed” to be your potential special some one who you might meet by accident walking down Spring Garden Road on a sunny, Sunday afternoon.
Note, however, that this census question does not take into account people who are living in a common-law relationship. StatCan was concerned with marital status as opposed to “relationship status” when asking this question. The number of common-law relationships in a tract therefore muddles the values because some one who is “never been married (single)” or “divorced,” for instance, may actually be living with some one in a common-law relationship. In the future I’ll manipulate the numbers to account for this, so for now understand that these maps, strictly speaking, reflect marital status in Halifax, Nova Scotia.
Population of Halifax, aged 15 or above: 312,650
Males, 15+: 148,390
Males 15+, not in a marital relationship: 74,490 (50.2%)
Females, 15+ 164,260
Females, 15+, not in a marital relationship: 90,350 (55.0%)
Please feel free to comment on the maps or to note any errors to be corrected. In the mean time, Happy Valentine’s Day.
Citations and disclaimers.
These maps were published with data gathered from Statistics Canada 2006 Census Tracts as well as from aggregated data retrieved from the Equinox data delivery system (Tables 97-552-XCB2006005 and 97-552-XCB2006006). This data was used strictly for scholarly research purposes and in no way in the pursuit of any commercial or income-generating venture.