Blog

Arc 5: Map Design with ArcGIS

Introduction

This exercise has three goals:  1) to review and reinforce GIS skill learned so far; 2) to put some of the design and symbolization principles from Monmonier’s Mapping it Out into practice; and 3) to learn how to create a map layout for print or export.

The example included here address how much and where land was plowed for crops in the 1930s.  It brings together maps based on two distinct spatial primary sources:  the 1935 Agricultural Census and a set of digitized aerial photos taken in Weld County, Colorado in 1938.

 

  1. Add data

All data are in the Workshop 5 folder.  For this exercise we’ll use two of the same datasets as in the previous Overlay Excercise.

  1. From the Design1.gdb geodatabase add the following polygon layers:
  • Weld_1930s (land use polygons digitized from a sample of 1938 air photos—8 cells of 9 square miles each)
  • AgCensus_1935 (land use attributes from the 1935 agricultural census for all Great Plains counties)
  1. Save your project by clicking the Save button, navigating to your Workshop 5 folder, and assigning an appropriate name. This will create a single file with the extension .mxd. Most of the finished maps you create for final output in ArcMap will be saved with an .mxd extension.  Each time you want to change a map design, you should immediately save it as an .mxd file with a new name, using the Save or Save As commands.  Today, you should click Save about every 10 minutes or so as you go along.
  1. Explore data
  1. Locate the small air photo polygons in relation to the much larger county map. Look at the attribute tables for the 2 polygon layers in order to understand the information they contain. For the AgCensus_1935 layer you can refer to 04254-0009-Codebook.pdf for variable descriptions.  Keep in mind the two very different scales represented by these two datasets.

III. Sybmolize data

  1. Make a map showing the amount of cropland in each county of the Great Plains. The best attribute for this map is CRP_XK_A. For the county-level AgCensus_1935 layer, choose an appropriate symbolization scheme.   (You may need to refer to Workshop I:  Mapping Great Plains Agriculture section V. Sybolize attribute data to refresh your memory.)  To start, right click on the AgCensus_1935 layer name in the Table of Contents and go to Properties and then the Sybmology tab.  For this map, the Quantities, Graduated colors options are most appropriate.  You’ll need to set the Value field to CRP_XK_A, select a color ramp you like, and choose a number of classes (then click Classify to determine the range of each class).  The result should be a map of counties in which those with the most cropland are darkest in color, and those with little cropland are lightest in color.
  1. Now zoom in to the Weld_1930s aerial photo layer, and sybmolize this map to show the locations of cropland in each cell. Again, you’ll right click on the Weld_1930s layer name in the Table of Contents, then select Properties and the Symbology tab. This time you’ll select Categories, Unique values.  Set the Value Field to Land_cover and click the Add All Values button below.  From this window you can set the color for each type of land cover identified from the air photos.  In this case we want to highlight the Cropland polygons only (perhaps using the same color as in your county map?), and set all other polygons to show a black outline with no fill.
  1. Separate layers into two data frames

Once you have both maps designed to your satisfaction, you’ll have to confront the fact that it will be very difficult to present these two maps at the same scale, since one covers only a small part of a single county and the other covers more than 1000 counties.  If you zoom the map so that one is visible, the other becomes hopelessly obsured.

One solution is to present the maps at two different scales on your layout.  In cartography, this approach is called an “inset map” or a “locator map.”

  1. So far in this class we’ve only employed one data frame, which in the Table of Contents is called Layers. Now it is time to branch out. Go to Insert and select New Data Frame.  Notice that in the Table of Contents the label New Data Frame has appeared in bold, and your maps have disappeared from the map window.  Right click on Layers and select Activate:  it becomes bold, New Data Frame is no longer bold, and your maps have reappeared.  In Data View, you can only see and work with one data frame at a time.

You can create as many data frames as you want, and when you Add Data to insert layers, the new maps, tables, or images will appear only in the currently active data frame (although it is easy to then drag them to other data frames if you want the same map in more than one).

  1. Rename the two data frames. Right click on Layers, select Properties, and then the General tab. Change the Name to read “Weld County Cropland.”  Now do the same to change the name of New Data Frame to read “Great Plains Cropland.”
  1. Click on the AgCensus_1935 layer and drag it down to the Great Plains Cropland data frame. A duplicate map layer appears with all of your symbolization intact. We don’t need the original any more, so right click on the original AgCensus_1935 layer and select Remove to delete it from the Weld County Cropland data frame.
  1. Zoom to the full extent of both maps. Activate Weld County Cropland then click the Zoom to Full Extent button. Activate the other data frame and do the same.

You’ve now put your two maps in separate data frames, and selected a scale that is appropiate to each.  Great—but now you can no longer view the two maps at the same time.  Because we are working in Data View, we can only deal with one data frame and scale at a time.

  1. Create a layout
  1. Now it is time to discover a section of ArcMap we haven’t dealt with yet: the Layout View. So far, all of our exercizes have taken place in Data View, which is the nuts-and-bolts mechanical part of ArcMap, where the interesting GIS analysis and basic symbolization takes place.  But once the hard core analysis is done and the map is symbolized to your satisfaction, we need more artistic options, and that’s where Layout View comes in.

 

Look in the map window section of the ArcMap interface, and go down to the bottom left corner.  Two tiny icons to the left of the scroll bar allow you to toggle between Data View and Layout View.  First, hold your mouse over each to confirm that you’ve found the right buttons—they’re not easy to spot.  Then click on Layout View.

 

Now you see a mocked-up page with your two maps both visible, though probably not where you want them to appear.  Layout View mimicks graphics programs like Adobe Illustrator (although with many fewer functions), so you might find commands that are familiar from that environment.  Here you can do the final design for an output map.

 

Key concept:  Complete all analysis and map symbolization in Data View before you start working in Layout View.  Going back later to change color choices, data categories, or the scale will often create havoc in Layout View, requiring re-working your final map design.  After years of heartache, take this advice from a grizzled veteran!

 

Note:  the program defaults to a Portrait layout, but you can change to Landscape if you like.  It also defaults to an 8.5 x 11 inch page size, but this can be re-set.  Make both changes by going to File and selecting Page and Print Setup.

 

  1. Click on each map to highlight its window, which can be dragged to a new location or resized by clicking on an anchor around the outside. Arrange the 2 maps as you like on the page. If you think that the Weld_1930s maps are still too hard to see, you could consider going back into Data View and zooming in on only one of the cells.  Then you would come back to Layout View to complete the map.

 

  1. From the Insert menu, insert each of the following map elements. In each case, use your own design sensibilities to choose from among the many options offered.

 

  • Title (create one)
  • Scale bar (one for each map, since they are at different scales)
  • Text box (in which you should identify the map author and date)
  • North arrow
  • Legend (one for each map, unless you used the same colors for both)

 

You can drag these elements into correct position, resize them, and format text (double click on the element, click the Text tab, and click Change Symbol) as you like.  If you don’t like the outcome, just select the element, hit delete, and start again.

 

Warning:  While it is very helpful to zoom in and out on the page and pan around as you do the design work, be careful to use only the zoom and pan buttons for the Page View (and not the very similar buttons connected to Data View).

 

 

  1. Generate a final map

 

  1. Once you are satisfied with your map layout, you can create two kinds of output. First, print a hard copy, by going to File and selecting Print. In order to get full color, send it to HP Color LaserJet CP4005 PCL6, which is the color laser printer in Kirk 213.

 

  1. Then export a .jpg file of your map, which could be used to post to a web site or to insert into a word document. Go to File and select Export Map. From the dialog box, set Save as Type to JPEG.  Here you can also adjust the resolution if you like (many publishers specifiy 300 dpi for print graphics, while web designers like 150 dpi or even lower for online purposes).  You can save the file to your Workshop 5 folder, then copy and paste it into a Word document.
Advertisements

Using Google Earth as an entry into basic GIS

By Daniel Macfarlane and Jim Clifford
From the Otter.
In the first installation of this series, we discussed how to use basic Google Maps to create a custom map. You might be surprised to know that, in doing so, you are actually doing GIS mapping (Global Information System – basically, GIS involves merging of cartography, statistical analysis and database technology). I’ll admit I didn’t know that. I asked how I could turn my custom map into a GIS map; Jim told me I already had. So if you have used the custom maps function provided by Google, you are already using a basic form of GIS mapping!

In this installation, we will discuss Google Earth, another free and easy option for creating a basic GIS database. Using Google Earth we created a multi-dimensional form of Dan’s original St. Lawrence Seaway and Power Project map.

Converting a custom Google Map to Google Earth is very easy: on the top of a custom map, just click on “View in Google Earth.” A prompt will ask whether you want to open or download the file. Unlike a custom map, which remains stored and saved online, a Google Earth map can be downloaded to your desktop (You can download Dan’s map. You’ll need a copy of Google Earth first, which is free).

Within a few seconds, you will be looking at your initial map in a three-dimensional space in Google Earth. All the same landmarks and such you already created will still be there. You now have a variety of options. For example, you can now tilt so that you are looking at the earth from an angle, rather than from straight above (like Google Maps, one has the option of “Street View” which simulates the map’s appearance form the perspective of someone standing in front of a landmark). This is a particularly effective way of following the seaway route – start at one end, and go along from about a 45 degree angle.

Google Earth is a good place to start, as it is effectively GIS-lite software. It provides the option to create custom lines, polygons and points (see the buttons along the top of the map). In GIS these shapes are collectively know as “vectors” and they are one of the main ways we represent geographic features in historical GIS. We trace the roads or railways of old maps using lines, the outlines of building or agricultural fields using polygons and identify particular places or things using points. Google Earth also allows you to attach descriptions to these lines, polygons and points. In the map below, (download sample here) we’ve added descriptions to a few geographic features of Toronto. This is another key functionality of GIS, the ability to link spatial data with other forms of data. Google Earth limits you to names and basic descriptions and does not allow you to link statistical data or dates, but it still provides a lot of opportunity to build a historical GIS database of the landscape you study.

Toronto Waterfront GIS

There is a Historical Imagery” option which will be especially appealing for environmental historians and geographers; however, it is limited to what Google has inn its database (which in the case of the St. Lawrence Valley, only goes back to 1997). In the region Jim studies, the Lower Lea Valley in East London, Google Earth has a 1945 aerial map.

some_text

Another options is “layers”: a huge collection of GIS data included with Google Earth. You’ll find these layers located along the sidebar, you have the choice of adding thousands of provided options. Take if from us, you don’t want to turn these all on at the same time, or you will be faced with an overwhelming and jumbled map. These include normal features such as labels, places, roads, pictures, etc., but also allow NASA, National Geographic, etc. These are effectively GIS layers provided by Google and partner organizations and the process of creating maps by adding and removing various layer options will prepare you for how more complex GIS software functions.

The most useful for historical research and teaching is the David Rumsey Historical Maps collection (found under the gallery tab) that includes dozens of historical maps “pined” onto their location on the digital globe. However, the user is limited to what is already provided and creating more layers, historical imagery, and such that are specific to one’s project requires the use of more specific GIS software, which will be the subject of the next series of post. There are a lot of options for the United States, but the options for Canada are a bit sparse.

some_text

In the meantime, if you are a student, staff or faculty member at a university you should ask your map librarian for a trial copy of ArcGIS and/or download and install Quantum GIS (a less powerful, but free, open source GIS package).

Learning GIS: A Google Seaway Map

By Dan Macfarlane and Jim Clifford
From the Otter.

This is the first of series exploring my metamorphosis, under the tutelage of Jim Clifford, from a digital mapping neophyte to a master mapper … or, to appropriate a title from quintessential Cold War film, this could be called “Dr. Macfarlane or: How I Learned to Stop Worrying and Love GIS Maps.”

I did my dissertation on the history of the St. Lawrence Seaway and Power Project and, as one might expect, encountered a great deal of maps and engineering studies along the way. I wished I could somehow usefully bring together or juxtapose all this spatial and geographical information, not only for my own benefit but for others who might be interested. I had used Google Maps many times, albeit for run-of-the-mill purposes, and had even made a few custom maps (e.g. to give directions to guests at our wedding) but the type of a multi-layered and interactive mapping I envisioned for my St. Lawrence research seemed well beyond my capabilities. Throw in the demands of a new baby and a dissertation to finish, and the idea was effectively shelved.

But as I saw others in NiCHE producing excellent GIS mapping projects, I kept wondering if I could do it. With some down time after my defence I randomly opened Google Maps one afternoon and started exploring the custom maps feature. Pretty soon I had marked off the main seaway channel, and quickly found that I could use anchors and signs to designate key points of interests and different features.

Over the following few days, I came back to this custom map when I had a few moments, and began to add text, maps, and pictures to the various points of interest. Then I started tracking the lines of older canals which the seaway had replaced. In just a few combined hours, I had essentially produced the map you see in this posting. I conversed with Jim and Sean Kheraj about the map, and they gave me encouragement and advice (such as viewing this map in Google Earth, which will be the subject of the next posting in this series).

The whole process was quick and user-friendly, and done from my couch with various football and hockey games going on in the background. I am most certainly no computer expert – when someone refers to java script I’m hoping they are going to serve me a cappuccino. I had taken a first-year computer science course over a decade ago but if I remember correctly, the major assignment was creating a website for a virtual golf course. And I forgot everything I had learned in that class anyway – I just remember typing a bunch of 0s and 1s. Bottom line: if you are computer literate enough to be reading this post, you can probably create a custom Google map.

In the next post we will show you how to make a map like Dan’s Seaway Map and explore how to view this map in Google Earth.

GIS and Time

By Jim Clifford

From the Otter.

One of the major weaknesses in using GIS for historical research are the limitations in showing change over time. GIS was designed with geography in mind and until recently historians needed to adapt the technology to meet our needs.

Generally this meant creating a series of maps to show change overtime or as Dan MacFarlane did last week, include labels identifying how different layers represent different time periods (see MacFarlane Map2). More recently, ArcGIS and Quantum GIS introduced features to recognize a time field in data and make it possible to include a time-line slider bar or animate the time series data in a video.

UK Tallow Imports, 1865-1904 from Jim Clifford on Vimeo.

I am currently studying the sources of raw material imported into Britain during the nineteenth century as a way to connect the environmental consequences of manufacturing in London with the ecological transformations taking place in other parts of the world. One of the first industries I’m studying is soap making. These factories relied on an increasingly global supply of fat (tallow, palm oil, coconut oil, and cottonseed oil), potash, essential (i.e. fragrant) oils, and pine resin. As a part of this research I’ve started a small database tracing the source of some of these products. Tallow is particularly interesting, as the Russians dominated the trade through to the 1860s, when large herds of cattle and sheep raised on American and Australasia grasslands drove down the price of tallow and led to the collapse of Russian exports to the UK. The most obvious way to visualize this data is to make a graph using the features included in Excel.

neoeuropevsrussia

I also experimented, using a slightly different dataset, with the Source Map platform, to create a series of maps showing change over time. These maps visualize the changing geography of Britain’s tallow supply, but it would be impractical to create or present fifty maps, so they only work to show the general trend using averaged data from three years in the middle of the decades:

British Tallow Imports 1864-1866(Total imports: £2,737,000)

British Tallow Imports 1874-1876 (Total imports: £2,407,000)

I wanted to see how I could best represent this data in GIS and I also wanted to learn the new time functions in ArcGIS. The result is the video (above), showing the changing source of tallow, based on quantity, from 1865 through to 1904. The advantage of this time series video is that it shows the volatility of Britain’s tallow supply, with significant booms and busts from each of the regions. Whether this particular animated map provides any insight not available from the Excel graph is up for debate, but the method might become essential as we create significantly larger text mined datasets in the Trading Consequence project.

For readers familiar with ArcGIS and who have a copy of version 10, I’ve include the Data File. QGIS has an experimental add-on called Time Manager, but it might not function properly for the nineteenth century yet. We will try and develop a lesson for this feature in the future, but for now I’ll provide instructions for people with copies of ArcGIS. The hardest part of the process was rearranging my data into a format that would work with the ArcGIS time function. First of all, the dates need to be listed in a column, not across the top row. It is possible to cut and past using Excel and switch your rows with your columns, by right clicking and using the “Special Paste” options. Don’t use “Date” as the field name as this might confuse the database. I’ve used “t” instead and included a simple list of years from 1865-1904 (the dates repeat for each location). The other fields I include are the quantity of tallow, a code for the geographic location and the latitude and longitude of where I want the quantities represented on the map.

Screenshot from 2012-12-17 11:45:49

You can use a spreadsheet program, like Excel, to manipulate this data, but I generally find it is best to save it as a CSV file before importing it into ArcGIS. Once you have a CSV file (or once you’ve downloaded my tallow_xy.csv file below), you can launch ArcGIS. It would be possible, and perhaps preferable, to create a database table for each location and avoid repeating the dates over and over again for each location, but that would make it more complicated.

Once you launch ArcGIS 10 you can import the CSV using the import XY data option. It should automatically identify the fields called x and y as the spatial data. It will then instruct you to export the data so the program will add a unique identifier number (making it a part of your database). Right click on the layer and chose Export Data. It will then give you the option to add this new layer to the map. If you are using my sample data, I would suggest choosing a simple world map as the base layer. The next step is to tell ArcGIS which field includes the time data. Double click on the layer and find the Time tab in the preferences menu. Choose “t” as the time field and if you click “calculate” it should give you the time range of 1865 to 1904. Do not click the “Display data cumulatively” check box. Next you will need to change the symbology of the layer to display the points as circles that represent the quantity of tallow from each place. I used proportional symbols, but you could also choose graduated symbols. Finally, you need to find the time slider menu button (it has a little clock face). Enable the time slider and it should allow you to slide through the years or click play to see an animation similar to the video above (play with the setting to slow down the animation).

Data File

The Geospatial Historian: Part 1, Niagara Falls

By Dan Macfarlane

From the Otter.

The modules are designed similar to Programming Historian lessons, and will potentially be offered there as well. After getting the necessary software that allowed me to run ArcGIS 10 on my Mac, taking online training modules made available for free through the Carleton University library, and collaborating with Jim and Josh, I have been available to develop these skills beyond what appeared in my original posts. What follows is the first of a series of examples of what will appear in our hands-on learning modules on using historical GIS.

The two maps contained in this post are both of Niagara Falls, and are part of my ongoing research about the engineering of the Niagara Falls hydro-electric landscape – basically, showing how Niagara Falls borders on the artificial – to the point that one would be tempted to say that the waterfall is little more than an elaborate faucet – because it has been so manipulated by Canada and the U.S. to produce hydro power and the effect on the actual cataract.

I’m writing a book on this subject, but for those who are curious and don’t want to wait several years for the book to be done, I have a chapter on Niagara in the upcoming environmental histories of southern Ontario collection that will coincide with ASEH Toronto (in fact, I’m co-leading a tour to Niagara as part of the conference) and I have an article on the topic forthcoming in the journal Environmental History. And for those who don’t want to wait several more months, see a blog post I did earlier this year on ActiveHistory.ca

Back to the mapping. For the first map (Hydro-electric Landscape of Niagara Falls) the initial step was finding the right base maps and layers (a base map is akin to a background image on which you do your digital mapping – in this case the base map is a Google Earth picture). I wasn’t entirely sure what I was looking for, but I knew that I wanted to create a map of the hydro-electric landscape of Niagara Falls. I scoured the internet for appropriate base maps of Niagara Falls, both present and historic. I also searched for aerial photos, which are sometimes used as base maps. In an Otter post earlier this year Josh gave useful tips on places to look for map data.

I used a .jpg of map from Canadian Geographic as a raster image, which are common to use as base maps (there are two main types of data/image: vectors, which are essentially made up of points, lines and polygons, and raster images, which like any digital image are made up of pixels on a grid), and then matched it up with an aerial photo using georeferencing (which I’ll explain a bit further down). At first I attempted a straight-above aerial view without any raster image – so that the different features you see in the final map were all on a white background. So after some quasi-successful starts, and going back for help to the various ArcGIS education resources I had at my disposal, I had made some headway.

The simplicity and non-cluttered effect of the blank background had some advantages, but I wanted a different angle as well as a representation of the surrounding communities geographic features, such as the water. I downloaded a map from Google Earth at an oblique angle to use as a base map, which is what you see in this post, and set out to georeference and digitize the Canadian Geographic map.

Georeferencing involves creating control points so that the software will match up the maps, as maps of the same area don’t necessary line up perfectly (some maps of the same features will be to such different dimensions and proportions that it won’t be possible to match them up without extreme distortion). So I put control points at Horseshoe Falls, Goat Island, Lake Ontario, Grand Island, various older hydro stations – and kept adding control points until the maps lined up.

Then I started digitizing features. Digitizing features really just means any type of drawing, tracing, or creating of something on a new layer as a digital vector (i.e. a point, line, or polygon). So the hydro station, the tunnels, the outline of the reservoirs are all digitized information that I created via the drawing feature on ArcGIS 10. An easy aid for doing this is to adjust the transparency on one of the two stacked maps so that you can see them both at the same time for tracing purposes.

All this of talk of layers might have you wondering what a layer is. Think of layers as transparencies on the overhead projectors your teacher might have used in elementary or high school. Each feature – be it road or a type of building or a type of tree – that you create is done on a new layer; thus, the information is independent of the other types of information, and you can add or remove it as you wish

To digitize vectors as a point line or polygon, one has to decide which best represents the real world feature. For representing cities on a world wide map, a point would be a good idea; but to represent the limits of that city in a zoomed-in view, a polygon would be useful. If we were looking at a map of North America, the Niagara River would be best represented with a line; but if we are looking at a close-up of the river, such as in the post, a polygon of the river’s shape would more effectively show its shape.

In the map you see here, the canals, conduits, and tunnels are line features, while the power stations and reservoirs are polygons (transparent on the inside to show the features in the base map, below). I didn’t actually use a polygon for the actual river and waterfalls, as I found that the the base map better illustrates those.

Adding labels was fairly straightforward, and it was relatively easy to shape text labels around certain features (e.g. running parallel to tunnels). For other features, there is an indicator line from the label text to the feature – this is a “call-out” line.

The second map (Modifications of Horseshoe Falls) shows the fill and excavations that were done in the 1950s, while the former crestlines indicate the amount of recession that has taken place. Because it used mostly the same approach and skills, as well as similar types of data (photographs and blueprints), I had used in creating Map 1, I completed this map much quicker.

Horseshoe Falls

There you have it. Nothing too complex compared to the kind of stuff GIS can do – in fact, what I was doing was mostly cartographic map-making rather than using the full abilities of the software to compute numbers and create new information (for example, the information in Map 2 could be used to measure the amount of feet the crestline has receded, or the amount of square feet that were reclaimed from the waterfall at the flanks). But I also found what I created quite useful for representing spatially what I was writing about in my research, and thus this is the type of end product that is within the reach of historians with a general competence around computers. Each map would benefit from further work, such as having a table of contents to better explain features, but I haven’t had time for that yet … after all, who knows what inane trend I’ve been missing while writing this post.