Blog #4 – Are We Still Talking About Cities?
Is she still going on about populated areas and their proximity to surface water? Yes. But this is the last one guys! I really want to follow through here, especially given our limited city dataset for the last couple of go-arounds. This time I’ll limit my technical explanations of the GIS processing part, since we’ve covered that pretty thoroughly by now.
I went looking for a Canadian dataset that had a more comprehensive list of all the populated places in Canada. The dataset we have been using lists 257 cities in Canada, while another dataset outlines that there are 6,129 cities in Canada. The dataset I found that lists these more than 6,000 cities? Behind a $500 paywall. I can’t blame them, the cost includes over 4.3 million populated locations around the world, last updated in December 2019. I continued on, I had at least 15 tabs open at this point: I went to the Open Government Portal and searched ‘Cities’ , I googled ‘Canadian Cities shapefile’ and found a limited dataset from the USGS , I found an awesome website called Canadian GIS and Geospatial Resources that could take me to all kinds of Canadian and Provincial data. Still no luck, but I kept noticing a dataset called ‘Geographical Names’ popping up that I wasn’t sure what it was, but I downloaded the spatial data from the Natural Resources Canada website, and it includes all Canadian populated places and so much more!
Seriously, this dataset is incredible, and you can choose to download it for specific provinces. Along with populated places, it also includes administrative areas (e.g. townships and Indigenous Reserves), vegetated areas (e.g. bogs and meadows), constructed features (e.g. portages, dams, and bridges), terrain features (e.g. peaks and lookout points), and water features (e.g. lakes and rivers). It should be noted that this is a point dataset, so larger areas such as administrative areas and water features are represented with a single point instead of a polygon. There are more than 350,000 points in this dataset, it almost covers the entire country with points!
You can learn a bit more about this data set at the Canadian GIS website .
I singled out the populated places from this Geographical Names dataset and made a new point shapefile, as a reminder, extracting features from a shapefile to make a new shapefile can be done by selecting the features of interest and choosing to export the selected features (Right click on feature à Data à Export Data à Export: Selected features). The features of interest can be selected manually in the attribute table by clicking the far-left column (see red star in the figure), or by using Select by Attributes.
Even within the populated places subset of data, there nearly 30,000 points in the dataset, absolutely bananas. Way more than the paywalled dataset we found before. The more detailed description of the populated places includes abandoned localities, administrative sectors, settlements, boroughs, cities, charter communities, community governments, dispersed rural communities, compact rural communities, villages, customs points, post offices, railway points, industrial areas, research stations, weather stations, and many others. We need to narrow down this dataset a lot, especially since some of these places, like post offices, are probably not what we are targeting with this query. Bringing out the Select by Attributes tool, we can select more of the “lived in” populated areas, like rural communities, villages, hamlets, and cities. The dataset also includes the names of communities located within cities, so these can be removed as well.
After removing post offices, railway stops, and other unrelated locations, the populated places dataset was reduced to 20,595 points, still a significant number compared to our previous dataset. We can run this sub-dataset through the same methodology we used in the previous posts to measure the distance from each populated place to the nearest water source using the Near tool.
What I found was: 2,174 places were more than 5 km away, 483 places were more than 10 km away, 31 places were more than 20 km away, and 2 places were more than 30 km away!
Of all these locations, 8 of them are cities: Camrose (AB), Humboldt (SK), Lambton Shores (ON), Martensville (SK), Melfort (SK), Steinbach (MB), Wetaskiwin (AB), Winkler (MB). There is some overlap with our previous findings (Camrose, Steinbach, and Wetaskiwin). Port Hope Simpson and Cartwright are considered towns and Bella Bella is considered an administrative area. Owen Sound is not included in this list because, although it is considered a city, its location shifted between the datasets. Its location in the new dataset (circled in red) is just over one kilometer from the river, whereas its previous location (circled in blue) was nearly 6 km away. This can happen in point datasets that represent a larger area, the representative location selected, usually a center point, can differ depending on the data source.
This means we should take our results with a grain of salt, and maybe put more reliance on the locations that are even further away from water sources. I’m particularly curious about the two locations that we found that were more than 30 km away from surface water sources: Dominion (YT) and Liebenthal (SK).
At a brief glance, I could not find much about Dominion, Yukon. Its size on Google Maps is quite small and it appears to be an operational gold mining site . Any settling in Dominion is likely driven by gold mining, and most people who work there likely live in Dawson further north, right along the Yukon River.
Liebenthal, Saskatchewan has a slightly different story. It is a very, very small Hamlet in southern Saskatchewan near the Alberta border. It has a church, a school, and a playground, and was established by German settlers in the 1900s, who were drawn to Saskatchewan for a number of reasons, including affordable and productive farmland. Despite its small size, it is a notable geographical name for the province and was a stop for the YouTube channel: Home Town Saskatchewan that visits the “little towns of the Canadian prairies”. Here’s his video on Liebenthal! (It’s quite windy there apparently).
I couldn’t find any information on whether groundwater is the main source of water in Liebenthal, but I know that it is commonplace in rural communities, so it is probably the same here.
I hope you enjoyed this example of a simple query we can answer with GIS! Believe me, it can get way more complex than this, but these kinds of simple queries can be used in a number of cases. For example, if you were looking to buy some property in a city, and you knew that you wanted to be close to a school, a grocery store, and a train station, you could use a similar methodology to answer your question! Next time, we will ask a different question, and start this process again. Tweet me if you have any ideas! Or any burning spatial questions.