Month: January 2021

Greetings,

This week I unfortunately had to skip the live lesson on Tuesday due to limited time. I was nervous about this “sacrifice” but had faith in managing with only the recorded lesson on my own. So the following day I sat down in my usual armchair, ready to tackle yesterday’s lesson. I had planned to have Arttu playing separately on my I pad, but for some reason the recording wouldn’t work on the tablet. This wasn’t however the biggest disadvantage as I was able to pause and alt-tab between the recording and Q-gis in peace and quiet. The rest of the lesson went smoothly as I just observed Arttu’s actions and repeated after him.

This week’s focus has been on map projections and their differences, as well as how one can go about portraying these differences visually. We also learned how to load data from a URL and not the local computer or a cloud service. Having done this, we went ahead measuring distances and areas, switching between projections to see how they differed. We mainly compared World Mercator’s projection and ETRS89 / TM35FIN. ETRS89 / TM35FIN is a projection that is specifically adjusted to project maps of Finland in the best way possible. World Mercator’s projection is a map that aims to project to entire world, which is much more challenging. Due to the way World Mercator’s projection is projected (along the equator) it leads to an increasing level of distortion of distances and area the further a location is from the equator. This could seem very impractical – and it is – but Mercator’s projection holds it’s strength in that it keeps all angles correct, making it the go to map for navigation.

The vast differences in area and distances between ETRS89 / TM35FIN and Mercator’s projection makes it anyway abundantly clear that Mercator’s projection should strictly be used for navigating purposes and that other, locally adjusted, maps should be used in almost all other cases. Sanna Korpi clearly displays this in her blogpost on this same matter (Sanna Korpi 2021) comparing how numbers on child population density (0-14 year/km²) differ between Lambert’s projection and Mercator’s projection. Mercator’s projection displays lower density overall with the lowest interwall being from 0-0,54 and highest interwall being 44,44-80,96 compared to Lambert’s 0-2,4 on the low end and 179,4-326,2 on the high end. This clearly shows that it’s not only the visuals that will differ when using different map projection. The calculations and answers will be faulty and hence distort the entire analysis and all conclusions based on that analysis. Conclusion: One should choose one’s  map projection with the uttermost care and consideration.

The main focus of this course lies however not on theoretical knowledge but on the practical skill of using Q-GIS. This week’s interactive exercise was to create a couple of maps and some tables. All of these was to display the differing levels of distortion between the projections. During the lesson Arttu goes through how to create a map that shows how many times bigger the area of Finland’s municipalities is on World Mercator’s projection compared to ETRS89 / TM35FIN. This process was clear enough and after having watched the recorded lesson I decided to play around with my map, trying to find the best way possible to communicate the phenomenon. I imagined it would be better to use World Mercator’s projection when displaying the distortion levels of that particular projection, so I chose to use it instead for my map. I also thought displaying the difference in area between the two projections in percent rather that “times bigger” would be better so I multiplied all values by 100. This was rather lucky as one of the homework assignments was to create a map displaying the differences in precent – two kebabs in one bite!

Picture 1. The map shows the levels of distortion on Mercator’s projection compared to ETRS89 / TM35FIN.

Something to note about this kind of map is that it visualizes the difference in distortion with a set of only a few colours (this case 7) and along municipality lines. The true case is that the level of distortion in Mercators projections slowly increase the further from the equator one looks. Since borders of municipalities seldom go in a strictly horizontal direction, we see some areas further south and north with in fact not ideal colours. A more realistic way to display the distortion would be to have a layer with small raster, gradually changing tone in colour.

The other parts of the home assignment we got was to do some measuring on a wider set of different projections. One set of measurement consisted of an area that makes up a “hat” on the head of “Suomen Neiti”, while the other part was measuring the width of Finland at about the latitude of the city of Vaasa. Here I also compare the different projections to an ellipsoid which truly gives accurate numbers on distances and areas. It’s not however possible to display the entire world in a 2D format on an ellipsoid projection. I display the following differences in distortion in the tables below (table 1 & 2).

In both tables we see the same set up with the same projections. We can see that ETRS89 / TM35FIN is by far the least distorting. Mercator’s projection is by far the worst on both measures while the Winkel Tripel- and the Loximuthal projections fair decently well. Gall Peter’s projection is an interesting case as it distorts rather much (256%) when considering areas, but with much less (155%) when looking at pure distances. All projections however distort more when considering area because of the nature of the calculations.

This week continues to bring just the right amount of work, not so much that one feels overwhelmed, but enough to feel challenging and meaningful! Although I couldn’t participate in the live lesson I still find that I understood everything and got all exercises done. Next week I think I will look at Monday’s lesson before our live lesson at Tuesday so that I already know beforehand should any problems occur. Last week I mostly “copied” Arttu’s exact work hastily putting together the map during the lesson, but this week when I had proper time and felt no stress, I indulged in experimenting with and exploring new techniques and methods. So if the first week was a bit shaky, this week was much more steady. What will the future bring?

Alexander Engelhardt

 

 

Sources:

 

Korpi. S. (2021) sakorpi’s blog (27.1.2021)

 

Hi!
I have chosen to write this blog in English although my mother tongue is Swedish and I understand Finnish as well, so that it would be useful to as wide an audience as possible. 

So, this course is something I have been waiting for ever since I heard about it from a friend who had it last year. There is something about maps that have always intrigued me and the thought of interactive maps filled with data ready to be played and tinkered with made me ecstatic!

The main focus of this course is to learn how to operate the software application Q-GIS which is an application with which one can produce and interact with data driven maps. During our first lesson we went through some key things about Q-GIS with our teacher Arttu Paarlahti continuously throwing in titbits of extra information here and there. At first we learned about the general set up and UI (user interface) in Q-GIS, but then Arttu made the important remark about any data management software – paraphrasing  – …it’s not very possible to learn how to use Q-GIS by just looking at it, you need to use it in practice to actually learn how to use it…

Naturally following we imported a set of data consisting of a map of northern and central Europe and the amounts of pollutive nitrogen in  the Baltic Sea for each country. First we just played around with the data; ordering around layers, changing colours, looking at data.

Then came the first actually interactive part of using Q-GIS. We were supposed to visualize the data about pollutive nitrogen in the Baltic Sea for each country by creating a choropleth map based on the available data. To achieve this we had to turn the absolute numbers into relative numbers which luckily was easy enough within Q-GIS itself. To complete the task we still needed to add a legend explaining colours and specifying numbers, add a scale bar and add an arrow pointing north. The map is shown below (Picture 1.).

Picture 1. The picture shows a map visualizing the relative amounts of pollutive nitrogen in The Baltic Sea for each country.

In my mind the map turned out quite decent and the procedure was clear enough as I just followed what Arttu did and asked questions when needed. Reading Sanna Jantunen’s blog (Jantunen 2021) lessened my like for the map as she pointed out some rather good points about her own map. She made a remark about the depth lines being redundant, only stealing attention from what is truly important. This I absolutely agree with!

For homework we got to produce a similar choropleth map on our own. This time I struggled quite hard as I had forgotten how to execute some of the actions which we used during the lesson. We were supposed to choose one section of the imported data about Finnish municipalities to visualize differences and similarities between the municipalities. This was pretty much the same exercise as we did during the lesson, but soon enough I got stuck. I tried and tried, but couldn’t find the function where you change the visuals of the map. At last I looked at the recording of the lesson and found the answer. What a relief! I really hope that these kinds of recordings don’t disappear when the distance learning is not the case anymore. What would be even better would be pre recorded theory lessons which the teacher would post on Moodle for students to look at before each live-lesson so that the live-lesson would then be as efficient and smooth as possible. This would save time for the teachers in the long run as well as demonstrating the dedication of The University of Helsinki to the wonderful possibilities that technology could add to education. 

The map which I created visualizes the geographical distribution of Swedish speaking Finns in Finland. As can be seen on the map (picture 2.) the Swedish speaking population is concentrated to the coastal areas in the south, to the region of Ostrobothnia and to the Finnish archipelago. To reflect the harsh lingual divide in Finland I chose to use a rather polarized range of red-ish colours to emphasize the differences between municipalities and different regions in Finland.

Picture 2. The Picture shows the relative amounts of Swedish speaking Finns in the municipalities of Finland.

Martta Huttunen made a similar map which she put in her blog (Huttunen 2021), but concerning the distribution of Saame speakers in Finland. Saame speakers are predominantly found in the northern parts of Finland, but interestingly  a small minority (1-4%) can also be found in Helsinki. Huttunen continues to explain the reasons behind this, a key factor being The University of Helsinki being the only university south of Oulu where one can study Saame. 

I have found the content of this first week quite motivating. Although the data we analyzed and visualized was rather dry and not highly interesting the exercises were challenging and I felt like I actually learned something new and important compared to many of the earlier courses I have had. I look forward to the weeks to come as we get to delve deeper into the world of Q-GIS and geoinformatics; failing and succeeding, struggling and overcoming, but foremost – learning and preparing us for a road to masters.

 

Alexander Engelhardt