Urban Heat Islands

[MadelineB]

Hello,
This is an interesting project. However, adding to the excellent suggestions by the expert in the previous post, and looking at your image plot, I wonder what your hypothesis is? Your hypothesis should help you determine which image graphs to show, and even the order that you present them.

From a graphical design point of view, I echo that expert’s suggestion that you consider showing separate side by side image plots for each of your selected layers (individual features) that you think might impact temperature …

There are several advantages to plotting the temperature map separately from the green space map and other environmental layers. First, that allows you to display a legend for the graduated color scale for the temperature map, using the appropriate color and range (e.g. orange to red) and then display a different legend and different color range for the green space (e.g. shades of green) and each of the other environmental features you are considering.  Second, it should be easier for you to describe the specific areas of interest and third, the science fair judges will find it easier to understand what you are focusing on.

With respect to your questions about settings and the formation of your “heat map”, your “images” are not viewable. Perhaps you could use screen grab and paste the image in a file as a png?

And, yes, you should probably include some of the actual table data and correlate that to the graduated color scheme you use for the corresponding image file, with a legend for the temps/color.

Again, like the expert, it would help us to help you if you could describe your hypothesis as well as what environmental features you consider might affect the temperature.

Thank you,
Madeline

[amyCC]

Hi - Madeline gave you some good input on next steps and on thinking about the visualizations you want to show in your paper.

I'm sorry that your images didn't upload. It's hard to tell what problem(s) you were having (or what settings you were trying to show).

I don't know what layers the one image we can see reflect. I see though what I believe are the transverse file dots that I am guessing relate to temp data? But do you have all three AF/AM/PM layers showing? Did you do any styling of the layers (using the QGIS options)?

Just seeing all those dots doesn't really suggest anything if you have all three times of day showing (and each is represented by a single color). (That's the default styling, so I saw it that way, too, when I set up my maps.)

1) I suggest you turn off the layers and focus on one time of day.
2) Then work with the styling for the layer. (You'll find the options by right-clicking on the layer, selecting properties, and then experimenting with the "Symbology" settings.)

You can use the "graduated" styling to use either a gradient or a spectrum that correlates to temperatures. This will make it easier to see how the temperatures vary in the location.

You can show a map showing the temperatures at one of the three time intervals, like this:

AF temps

uhi-afternoon-example-compare.png (2.33 MiB) Viewed 6817 times

If I turn off AF and look instead at the AM layer, it's different:

uhi-am-compare-example.png (2.16 MiB) Viewed 6817 times

And then you might, for example, turn on a tree canopy layer that then lets the viewer see where trees fall in relation to those temperatures:

uhi-am-example-trees.png (2.06 MiB) Viewed 6817 times

As you work with your layers and with the styling, you will be able to help viewers see the correlation(s) (if any) you are making between temperature and environmental and urban features (like trees, bodies of water, etc.).

What layers you have will impact what correlations you are able to suggest. So it depends on what you have pulled in.

Here's a quick look at the Symbology screen. (There are other options. This is just with one of the available options selected. It's important to work with these setting through because the default is a single color for ALL of your data.)

uhi-symobology.png (103.36 KiB) Viewed 6817 times

There is a lot more you can do in QGIS, but in the timeframe you have, I think you need to focus on creating your visualizations so that you can put them into your paper.

Good luck finishing up!

Amy

(Note - In looking back at my post, I see there is some weirdness in my legends that I overlooked when snapping the screenshots. Don't let that confuse you! Select your temp for the values to use for plotting the colors and creating the legend.)

[MadelineB]

Hello,

I want to emphasize the importance of following Amy’s suggestions about “styling” for each individual layer. And, I think it is good to clarify some of the terminology used in dealing with geospatial data:

(1) “graph” or “plot” typically mean the same. In your project, think of graph as the collection in one plot as a collection of “layers”, typically starting with a base layer (typically giving the geographical outline, and not necessarily including any details like roads, structures, waterways).

(2) in selecting a base layer, it is important to be careful to exclude extra detail for areas included in a graph for which you have no key data … e.g., try to include just a bare outline of any boroughs not included in your analysis. On the other hand, when you add color, be careful to show the surrounding water and watch out for how the software is showing areas within your target for which data (e.g., temperature) was not recorded.

While trying to clarify terminology, it would be helpful if you could explained what you mean by “heat map” and how that would fit in with your hypothesis and the layers you intend to show.

Your choice of individual layers should be based on your hypothesis. In a science fair, judges will pay particular attention to how thoroughly your describe the information about each layer. In addition to including a legend for each layer which usually describes the graduated color and clear labels for the axes, your description should include the date(s) represented by the data in this layer and what the this layer is showing in comparison to the other graphs and layers as well as a citation and link to source of the data.

In my previous post, I recommended that you use separate, individual graphs to display the key layers. Another major advantage of showing the key layers in separate individual graphs is then you have room to include a paragraph with the info in my paragraph above with each layer!

These paragraphs should provide your description of how each layer relates to your hypothesis and relate to the other key layers.

And, again, following Amy’s suggestions for paying attention to styling is really important … those details help you write detailed descriptions of each layer so the judges can follow your “story line”!

I realize I’m suggesting that you pay attention to a lot details ... but these details will help your viewers understand how your geospatial data relates to your hypothesis!

Good luck wrapping up this very interesting project!

Madeline

[ciascience00]

I'm sorry that the images didn't show up correctly, I'm not sure why that happened. But, I followed your advice on the symbology of my points. It took a little tinkering around, but I successfully stylized the points in various ways (I've attached the images as files this time, hopefully you can see them now!)

And thank you for the terminology advice, I wasn't really educated on that topic from before, and how I should apply it for my research. Speaking of which, my final research question + hypothesis is as follows:

Effects of Green Spaces on Impervious Surfaces, “How does the temperature surrounding impervious surfaces in a city, vary around a city’s green urban spaces?”

"If a certain green urban area, such as a park, is next to an abundance of impervious surfaces, then the temperature of such impervious surfaces will decrease/be lower compared to impervious surfaces that aren't surrounded by green urban areas."

For my 'story-line'/graph I think I'll mainly focus on what the temperature points display near green spaces, ex. how the color of the points vary around a park. Also, I didn't even think about how time could be included with this, (with the AF/AM/PM, and it took me a while to realize that AF meant afternoon ), so thank you for pointing that out! I'll definitely include it in the explanation in my paper.

Moving forward, I'll separate the different time layers into different graphs; that'll show the temperatures throughout the day. (Just for clarification, does AF = afternoon, AM = Morning, and PM = Evening/Night?) I'll use the heatmap stylization for an 'overview' visualization and I'll use the graduated stylization for a more in-depth look.

[ciascience00]

And here are the symbology settings I put for the graduated stylization. I think I'll lower the amount of classes though.
As for the heat-map stylization, I applied the same color ramp and tinkered around with the radius settings.

Quick question however, how would I make an outline of the boroughs I will include compared to the ones I WON'T include?

[amyCC]

You’ve made a lot of progress! The heatmap styling you put in place is very nice! What I’m not seeing, however, is how you are showing green spaces? In the example I showed above, I showed how we can look at temperatures — and how I can then turn on the tree canopy layer (shown as green triangles in my example). That lets the viewer see temperatures in relation to tree canopy to see if there are observable trends.

If your hypothesis is about green spaces, then your maps need to be able to show those. The heatmap of temperatures only has meaning in terms of what a heat island is if you are showing the surround spaces. I see you have a land cover raster layer that is greyscale. How did that work out for you? Did you pull down any specify data sets related to tree canopy or green spaces?

Also, did you consider water sources as potentially important in the temperatures for these areas? I think it’s a question you might be asked about. There are, of course, many different features that can be related. The types of buildings, too, play into this equation.

Yes about AF, AM, and PM. Whatever source you pulled from should have a readme file that explains that. I don’t know that this is the Brooklyn data you used, but as an example:

IMG_9681.jpeg (240.84 KiB) Viewed 6796 times

Also, if you used a data source from this site, you might find that there was also a PDF. It’s a really good idea to look through that. There is a lot of information shown and explained in the PDF summary for each location. For example, there is a pdf on this page https://osf.io/pzjn9

I’m not sure at this point if you need to have different images for the underlying map. Madeline may have suggestions for you. You do want to be able to clearly show the separate maps though — it’s hard to tell where the dividing lines are between them right now. You can just turn layers on and off to get views you need for screenshots.


Are you feeling like what you are seeing is supporting your hypothesis? I’m curious what conclusions you are drawing based on what you have mapped and what other layers you have brought in to show differences (if any) between these two locations that you can relate to the features you are looking at (like trees and water).

Amy

[MadelineB]

Hello and thanks for sharing your hypothesis! That help to focus our suggestions.

First, to answer your question “how would I make an outline of the boroughs I will include compared to the ones I WON'T include?” That will depend on the geographic area included in particular layer(s). I think you won’t need to show the boroughs you aren’t including as long as you don’t include layers that include them. The graph you showed us in your Feb 13 post seemed to include layers that included data for boroughs that you aren’t focusing on. Let me know if this answers your question!

I’m puzzled by your graph that seems to show the pathways used for measuring temperature. You label that graph “turbo-gradulated stylized points” but I do not see any gradations in the “gray scale” on the map, all I see is black. Also, since the info you are displaying is about temperature, I suggest using a color gradient from yellow (for the lowest temp) to bright red (for the highest temp), and NOT including blue in the gradient!

As I suggested yesterday, be sure to include titles and a legend on each graph that shows the temp values (F) and the color gradient and it would help the reviewers if you used the same range of temperatures for each map showing temperature, so that the brightest/hottest red represents the same temperature value across your various graphs. This means that the am and evening temperatures might not actually show any temperatures at the highest temperature seen in the afternoon.

Amy has made some great suggestions particularly about finding the read me file and the pdf report. I also agree with her suggestions to display the green spaces and bodies of water. I am concerned that trying to put several layers on the same graph might make the graph too busy. You might also find that info on one layer might obscure crucial info on previous layers. It might take more effort on your part, but the info might be easier to talk about for your reviewers if you have, say 3 graphs, each with its own color gradient: temp (yellow to red), green spaces (graduated green!) and bodies of water with a graduated blue.

For evaluating your hypothesis, I wonder how you plan to select the size/diameter of areas of impervious surfaces and the size/diameter of the green spaces?

Again, it’s clear you have made good progress! I’m curious when your final paper is due? You might need to decide on some key points to consider for this project … and maybe consider extending your research for a follow-up project for next year!

Good luck!
Madeline

[ciascience00]

I feel like the files I have now do already support my hypothesis. And as for the green spaces, I've just been using the labeled parks already shown in the OpenStreetMap. However, I considered your point of trees and such, so I also downloaded a CSV file of NYC Street Trees and added that into the graph.

I'm not exactly sure how to use the Land Cover Raster Layer for analyzation, so I've just been ignoring it for the time being. I'm not sure if it's really needed at this point, considering my current time frame. However, I could be wrong.

The attached file I've provided does show a water source, and a park, surrounded by temperatures, which I thought was good. Also, I will turn on and off the layers and screenshot my graphs separately. I'll organize them into my paper as "morning, afternoon, and evening"

As for the color scaling, I originally thought that the "Turbo" color ramp was good because the blues would represent lower (or just less warmer/colder) temperatures, while the red would represent the higher ones. (But now looking back at it, it's not zoomed in enough to see the various colors-) However, I do understand the point of 'not making the graph look too busy' so, is it alright if I just change the graph into "Red scale" color ramp instead? I think that may represent it better.

I'm not sure how to select the size/diameter of said impervious surfaces and green spaces. My plan was to utilize the OpenStreetMap's labeling for that, and to just use the 3 boroughs I'm examining as my size.

Also, how would I create a legend/title on my graph? Or are you referring to me labeling the legend/title when I'm writing my paper and describing my graph?

[ciascience00]

Okay, disregard what I said about the Tree Canopy layer. I tinkered around with the layer's symbology and transparency after you mentioned it and now it does add some more understanding to my graph.

[MadelineB]

Hi,

There are several videos if you google "how to display color gradient legend in qgis" ...

and the same for adding a descriptive title:
how to add title in qgis

As for size/diameter, that will depend on how you plan to make the comparisons between surfaces without surrounding green spaces, etc and those with. I suggest that you decide on size/diameter and be consistent across your comparisons.

Madeline

[MadelineB]

More thoughts on "size/diameter" ...

You ask: "I'm not sure how to select the size/diameter of said impervious surfaces and green spaces. My plan was to utilize the OpenStreetMap's labeling for that, and to just use the 3 boroughs I'm examining as my size."

The reason I asked about "size/diameter" for your comparisons is that I got the impression that your plan was to look for several "paired" subsets within the 3 boroughs ... comparing the observed temp for a selected area of impervious surface without surrounding green, etc to the observed temp for a selected area of impervious surface with surrounding green, etc. I must have not understood how you planned to do the comparisons.

Madeline

[ciascience00]

Hello again, I finished writing my paper on Saturday. It wasn't exactly perfect, but it was still pretty good. I'd just like to thank you for all of your help, because I probably could not have done it without any of you. Please let me know if you need anything else regarding this topic.

[amyCC]

Thanks so much for letting us know that you did finish and get it turned in. I'm glad to hear that!

I would love to follow up with you, if you are willing, with a few questions about your project. (I often write stories that highlight student projects.) If that's okay, I'll reach out by email. (I have your address, so you don't need to post it here.)

You did a good job sticking with your project even when it became clear that it wouldn't work exactly the way you initially expected.

Amy
Science Buddies

[MadelineB]

I want to add my thanks to Amy's thanks and congratulations on finished your paper. That must give you a nice sense of accomplishment!

Best of luck for your future research!

Madeline