Draft, 19 September 2016
This page should be in a useful state, but still needs work before it's finished.
Many years ago now I was involved in a project called Mapping Our Anzacs at the National Archives of Australia. The plan was to build a map interface to 376,000 digitised WWI service records that enabled visitors to find records by the place of birth or enlistment of the service person. To do this I had to extract placenames from the file titles, find the coordinates of the places, and then figure out a way of displaying them on a Google map. We ended up with many thousands of places, and given the limitations of browsers at the time, it wasn’t easy to put them on a map. If you tried to display them all at once everything would slow to a crawl. But still…
I remember the first time I got all those places to display – no fancy interfaces or animations, just thousands of markers obliterating much of Australia. Thousands of places, large and small, that had sent their young people to war. Just seeing those markers crammed in the browser window was a powerful expression of the impact of the war on Australia.
I’ve still got one of the original data files, so you can try this for yourself. You’ll need to install Google Earth. Then just download the anzacs.kml file. Once it’s downloaded, just double click on the file to open it in Google Earth. You might need to navigate your way round the globe to find Australia, once you do, zoom in, and in. You might also like to have a look at the UK.
Even simple maps can carry a powerful message.
Here’s a pretty standard map of the world. Have a close look at it and answer one simple question – which is bigger, Australia or Greenland?
Judging from the map alone you’d think that Greenland was considerably bigger, but it’s not. The land area of Australia is 7,692,024 km2, while Greenland is only 2,166,086 km2 – it’s less that a third the size!
So what’s going on? As you probably know, there’s no simple way of displaying the surface of a globe in two dimensions. In fact cartographers have created many alternative projections, all with their own benefits and limitations. You can browse a list of different projections on the Radical Cartography site (you might need to click on ‘Wall maps of the world’). Note the projections that are labelled ‘Equal area’ – they represent the area of land masses accurately, though the shapes may be distorted.
The projection above is the Mercator projection, created way back in the 16th century to help sailors navigate the world. Despite its distortions, the Mercator projection is still widely used and, most importantly for anyone embarking on a digital mapping project, it forms the basis for just about all online mapping services, such as Google maps.
We put a lot of trust in online mapping applications. We believe them when they tell us to take the next exit on the freeway. We expect them to tell us the ‘truth’, but like other forms of visualisation, there are many assumptions built into the creation of maps. Maps can lie.
You may have seen a story in the media recently about a family in rural Kansas that have been accused of all manner of computer-related crimes. Why? Because investigators on the hunt for spammers and scammers try to find them by geolocating their IP addresses (the network address your computer uses to access the net). But when geolocation services don’t have enough data for a precise fix, they just return an address in the geographic centre of the country – and the centre of the USA just happens to be a farm in Kansas. People just assumed that geolocation services were telling the truth.
So once again we have reason to be sceptical of online services. But once again digital technologies can help us to see things differently, to break free of Mercator’s grid.
The True Size Of… is a fantastic digital project that compares the standard web Mercator projection with the real size of countries. It’s easy and fun to use:
Right click on any of the coloured outlines to remove them from the map.
Enter ‘Australia’ in the search box.
Once the coloured outline appears, drag it towards the north pole. What happens?
Try layering Australia over Europe, or over Greenland!
Let’s start off by creating a simple map with Google’s My Maps. You’ll need to have a Google account to create and share maps.
To open up the My Maps editor you can either:
go to Google My Maps, and then click Create a New Map;
or go to your Google Drive account and choose New > More > Google My Maps.
Add a point to your new map:
Click on the marker icon (the cursor will change from a hand to crosshairs)
Click anywhere on the map.
Add a title and description in the info box. Click save.
Search for a place to add to your map:
Type a query in the search box to find a place – try ‘University of Canberra’.
The map will recenter and a marker will appear.
To add this marker to your map, click ‘Add to map’.
Click on the pencil icon to edit the content of the info box. You can also add a photo by clicking on the camera icon.
What if you already have data about places that you want to add to your map? It’s easy to import data from a spreadsheet by creating a new layer. We’re going to create a map of places on the Australian National Heritage List using a dataset from data.gov.au.
Download the CSV file.
Open up the CSV file and have a look at its contents.
One thing you might notice is that although there are addresses for each site, there are no coordinates – latitudes and longitudes. How will Google Maps know where to put the markers? As you’ll see, Google Maps (and Google’s Fusion Tables) will automatically attempt to turn addresses into coordinates. This process is known as geocoding or geolocation. We’ll talk more about this below.
In the My Maps editor click Add layer. A new layer will appear in the dialogue box.
Click Import under your new layer.
Select and upload the heritage places CSV file.
You’ll then be asked to choose the fields that will enable Google to position your markers. Scroll down the list and check ‘address’. If you want to see an example of the contents of each field, just click on the little question mark.
Now you need to select the field that contains the titles of your markers. Check ‘name’.
Yay! A map appears with lots of little markers. Try clicking on the markers. Zoom in to see things more clearly.
But wait a minute – you might have noticed that there’s a message in the layer details saying that some of the places couldn’t be located. Oh no!
Click on the Open data table link under the warning to see what the problem is.
At the top of the table you should see a number of rows highlighted in red. For some reason the
address field didn’t provide enough information.
Find the row for ‘Mawsons Huts’. You’ll see that the
address has ‘EXT’ instead of a state (presumably meaning ‘external territory’).
Click on the
address field and change the ‘EXT’ to ‘Antarctica’. What happens?
See if you can find any other obvious problems and correct them. If we were doing this properly we’d obviously want to make sure that all our places appeared on the map.
Close the data table.
Let’s try styling our map a bit to show some more information. Click Uniform style in the heritage places layer.
In the ‘Group places by’ dropdown list choose ‘class’. You’ll see the markers will now have different colours representing the three classes – ‘Historic’, ‘Natural’, and ‘Indigenous’.
This is a simple example of how maps can represent not only locations, but data relating to those locations. My Maps abilities to display data are limited, so for more complex datasets you’re probably better off using something like Carto (see below).
It seems that large areas of Canberra have no bus stops! Why? It’s just because My Maps will only display 2000 points at a time. Once again – beware of what online services hide!
Like the charts and visualisations we saw last week, Google’s maps can be easily shared and embedded. Click on Share to change who can see your map and create a public link. Click on the three little dots next to the map’s title to find the embed and export options.
Increasingly maps are presented as ways of exploring complex datasets. Have a look at the series of maps created by the American Panorama project, such as The Forced Migration of Enslaved Peoples. Examine the types of data being presented and the way the different visualisations interact. Excercise the critical skills we discussed last week – What’s missing? What’s unclear? What do you learn?
Another interesting example of data-driven maps is Orbis. This is an attempt to represent the experience of travel in the world of Ancient Rome. By adjusting the settings you can see how this experience changed depending on how much money you had, the transport options available, even the time of year. In this case the data is not really obvious, but clearly the map is based on much detailed research.
For more examples, ideas and inspiration have a browse through the Carto gallery.
Depending on the source of your data you may or may not have coordinates for the places you want to map. As we saw, My Maps (and Fusion Tables) will automatically try to geolocate address data. However, there doesn’t seem to be any way of getting the geolocated data back out of Google.
There are a large number of geolocation services about – most are commercial, though some will offer a free or trial service that will allow you to locate a limited number of points. Mappify is an Australian service that seems to offer a reasonable free allowance. Carto, which we’ll explore below, only lets you geocode 100 points a month before you have to start paying.
If you’re willing to make use of geocoding APIs there are more options available. You might have noticed that the OpenRefine tutorial we worked on a few weeks back includes a section on geocoding data. There’s also a wholly free and open source geocoding API called Nominatim made available by the Open Street Map community.
To introduce Carto (formerly CartoDB) we’re going to work through the Getting started with CartoDB tutorial. This shows you how to import data, create maps, and explore a variety of options for representing your data.
Once we’ve covered the basics, it’s time to try some more advanced techniques in the tutorial Merging data sources in CartoDB.
If you want to play around a bit more with the possibilities of Carto, you could try loading the anzacs.kml file. How might you represent this data?
So far the base maps we’ve been using are just the standard modern web maps we use every day. But if we’re using historical data we might want to position our markers on a historical map. There are plenty of historical maps around, the problem is their coordinates might not match those of modern geospatial systems. The way we get around this is by georectification. Put simply, we match known points between the historical maps and modern maps and then distort or align the historical maps so that they can be used with modern geospatial systems.
Have a look at the NYPL Map Warper to browse a large collection of georectified historical maps.
We’re going to make use of a public version of the NYPL Map Warper tool to georectify some Australian maps.
Go to MapWarper
We’re going to georectify this 1856 map of Melbourne.
Click through to the digitised version and then click on the ‘Download’ icon.
This map is available as a high resolution tif. This would be the best version to use as it would allow us to zoom right in. But it’s 112mb, so to speed things up let’s use the jpg for now. Select the ‘JPG’ option and click Start download. You’ll need to unzip the downloaded file.
Go to MapWarper, login and click on the ‘Upload Map’ tab.
Copy the basic bibliographic details from Trove to the MapWarper form – title, publisher, date etc.
Click on Choose file button to select the map we just downloaded.
Once the form is complete, click on Create.
Once the map has loaded click on the ‘Rectify’ tab. You’ll see the map you loaded on one side, and a modern map on the other.
Pan and zoom the modern map until you’ve found the location of your historical map.
Now the fun begins! Identify points that you can see on both maps (such as the corners of streets).
Click on the pencil icon to ‘Add a control point’ to your historical map. Just click on the map at the point you identified.
Repeat this process for the modern map.
Once you’re satisfied that you’ve identified a pair of matching points, click the Add Control Point.
Repeat this at least three times with additional points.
Once you’ve added your points, click the Warp image! button.
Once it’s done, click on the ‘Preview’ tab to view the results!
Now try on your own, here’s a few searches that will find digitised maps in Trove:
Find, upload and geocode your own map and share the results on Slack.
How can you use these georectified maps? In the bottom left hand corner of your Carto maps you might have noticed an option to change the basemap.
Neatline is another mapping tool that allows you to incorporate a range of sources and use maps to tell complex and uncertain stories. See their demos for some fascinating examples of what’s possible.
There are a number of tools available that enable you to create narratives using maps:
You can even use your georectified maps in StoryMap to narrate a journey around your historical map!
There are a number of timeline tools around, some free, some paid: