Quantum GIS Tutorial for Science On a Sphere®

Written by Audrey Randall

Summer 2014

Quantum GIS (or QGIS, or Quantum Geographic Information System) is a “free,

open source geographic information system.” It provides a way for non-programmers to
create maps using GIS information.

This tutorial will walk you through the basics of using QGIS to create SOS-
formatted maps, by going through the creation of the Earthquakes and Nuclear Power
Plants dataset, the Marine Life dataset, the Global Statistics dataset, and a few maps that
are not currently part of the SOS collection.

Further resources: An excellent set of general tutorials can be found here:

http://www.qgistutorials.com/en/. Thank you to Ujaval Gandhi for providing these to
QGIS users free of charge. Additionally, SOS’s Content Creation guide may be found
here: http://sos.noaa.gov/_downloads/docs/sos-content-creation-guidelines.pdf. Make
sure you read it BEFORE creating your map. There is a lot of information in there that
can save you from having to redo an entire project.

A couple of notes for users: First, QGIS can be finicky. Please SAVE YOUR
PROJECT OFTEN. Second, QGIS often has trouble exporting very high resolution
images. The SOS team recommends that users use the lowest resolution images that still
look good on the sphere— in most cases, that means a resolution of 2048x1024 pixels
for simple maps and 4096x2048 pixels for maps with intricate details.

Links for all the data used in this tutorial are provided. In cases where the links
are no longer valid, a copy of all the files used is available here:
http://sos.noaa.gov/Docs/qgis_tutorial_files.zip

1
 Table of Contents
Overview…………………………………………………………………………….4
Requirements of the SOS format…………………………………………………… 5
Some background information before you start
Vector vs. Raster Data
Downloading QGIS…………………………………………………………………6
Making a Base Map with Vector Data………………………………………………6
Making a Base Map with Raster Data………………………………………………6
Downloading Correctly Formatted Images
World Files
Importing Raster Data
Re-ordering Layers
*Example used: SOS’s Blue Marble image
Finding/Manipulating Point Data……………………………………………………11
Splitting a Column: A common format issue
*Example Used: WNO’s map of nuclear reactor locations
Formatting Point Data……………………………………………………………….14
Tab-delimited files in QGIS
Attribute tables (tables of data associated with layers)
Null values in attribute tables
Coloring points based on associated data
Changing point color, shape, transparency, and border
*Example used: NOAA’s database of significant earthquakes
Making a Map with Regions Colored by Data………………………………………20
Option One: Shapefiles on the Web…………………………………………..20
Example Used: Life Expectancy from Atlas of the Biosphere
Option Two: Shapefiles from Spreadsheets………………………………......25
*Example Used: ITU’s database of landlines per country
*Software Used: CartoDB.com
Option Three: Adding Attribute Data to Existing Shapefiles…………………33
*Example Used: Life Expectancy from Atlas of the Biosphere,
Language Spreadsheet from Brown University
Exporting Maps as Images……………………………………………………………37
*Example used: Reactor Sites and Earthquakes Map
Legends, Color Bars, and Scales……………………………………………………...40
Exporting Maps as Layers…………………………………………………………….41
Creating layers based on data in the attribute table
Copying layer styles
*Example used: Reactor Sites and Earthquakes Map
2
Making Image Backgrounds Transparent ………………………………………….......44
Batch-editing images
*Software used: GIMP, free image-editing software for all platforms
Connecting Dots using the Points to Paths Plugin……………………………………...48
This is a useful tool for converting locations of something over time into a path.
For example, use this plugin to convert an animal’s location over time into an
approximation of the path the animal followed.

3
An Overview of QGIS

QGIS, according to its website, is “a user friendly Open Source Geographic

Information System (GIS) licensed under the GNU General Public License.” To draw an
analogy, it is to making maps what Adobe Photoshop is to editing images. It gives you
the ability to create maps from scratch or from imported data, or modify and add to
existing maps. The software is set up so that each imported map or map feature is a
“layer.” Layers can be made of point locations, background images, polygons such as
country borders, or more complex types of vector data, like WMS maps or ESRI
shapefiles. While QGIS can manipulate maps, it cannot create data. Users must do that
themselves, by manually creating a table of information or (more commonly) importing
data from the Web or other sources. The two main categories of maps that QGIS can
assemble, along with the type of data required to make them, are below:

• Maps of event locations, such as lightning strikes, earthquakes, tornados, or the

locations of tagged animals. These maps require importing a Comma Separated
Variable (.csv) file or a text (.txt) file. CSV files are simply spreadsheets that have
been saved in a form a computer can easily understand. Contrary to the name,
they can be saved with tabs, spaces, or other characters separating the
information, not just commas. If you are unsure how your CSV file is organized,
simply open it in a text editor such as Notepad to determine what characters
separate rows and columns. The information in a CSV file that could be used to
create these maps would have to include separated latitude and longitude
columns. If the coordinates are in the same column, see the section labeled
“Finding/Manipulating Point Data.”
• Maps of regions colored according to data, such as a map of countries shaded by
population. To create this map, you’ll need an ESRI shapefile that includes
country borders and names or IDs, as well as a CSV file that has country
population by country name or country ID.

Usually, maps are made up of a base map (the world map, either an image or a vector
map made of polygons such as continents) and overlaying layers of interesting data,
such as locations of phenomena. Base maps can be created with vector or raster data.
Vector layers can be scaled indefinitely, because they are made up of mathematical
algorithms that tell the computer where to draw lines and shapes. Every time you zoom
in, the shape just gets redrawn. Raster layers are images. They are composed of pixels,
so will lose quality the more they are stretched or zoomed in on.

4
The SOS Format's Requirements

Because of SOS’s high resolution requirements, it is generally easier to have a

base map of vector data, which can be scaled indefinitely. Raster, or image, data works if
it is high enough resolution, but images of sufficiently high resolution are uncommon.

Additionally, SOS's software needs to be given a specific map projection in order

to render the map correctly on the sphere. In QGIS, this is the default projection: the
Equatorial Cylindrical Equidistant projection, also known as WGS 84 or EPSG:4326.
While this is a relatively common map projection, it is NOT the one used by anything
that Google Maps has made, or anything with Google Maps as a background. There are
a lot of tutorials for QGIS that assume a Google Map will be a fine base map, so they
tell users how to import that and skip over the instructions for importing anything else.

Because of this limitation, the underlying map layer is a little more complicated
when making an SOS map than it is when making a normal map. Keep a few tips in
mind:

• It is certainly possible to import a vector map in the correct projection. They're

just not very easy to find. A good place to go is www.naturalearthdata.com. Click
on the green “Get the Data” button near the top of the page. From that page, you
can download the “Natural Earth quick start kit” or browse the available
resources. If you choose to download the quick start kit, you will get a zip file
that contains a folder called “packages”. Inside is a .qgs file called
“Natural_Earth_quick_start_for_QGIS.” To get started, it's a fun one to play with.
• If you want to find a vector map from another source, QGIS accepts the following
forms of vector data:
o KML
o SVG
o GeoTIFF
o GeoJSON
o PostGIS
o SpatiaLite
o MSSQL
o WMS/WMTS
o WCS
o WFS
Start by trying to find data in one of these formats. Websites from NASA, NOAA,
and other government agencies will occasionally have mapping data available to
the public.

5
 • It is also possible to put an image as the base of your map, provided the image is
of high enough resolution (2048x1024 or 4096x2048 pixels for SOS). SOS's FTP
server (ftp://public.sos.noaa.gov/) is a good place to find high resolution images
that are already in the right format.

Downloading QGIS

QGIS releases new versions often. The version this tutorial is written for is QGIS
2.2, which is no longer the current version. The SOS team recommends that you
download QGIS 2.2, which is available at http://www.qgis.org/downloads/. Find the
download link that works for your operating system.

Once you’ve clicked the link, the QGIS Setup Wizard should appear. Follow its
instructions to download QGIS 2.2. When you click “Finish,” several icons should have
appeared on your desktop with a yellow Q on them. You want to click on the one called
“QGIS Desktop” to open the most sophisticated version of QGIS. QGIS Browser is
simply a less versatile version of QGIS Desktop. If these icons do not appear, search for
“QGIS Desktop” in the search bar in the Start menu.

Making a Base Map with Vector Data

To insert a vector layer and style it, please see this highly helpful tutorial:
http://www.qgistutorials.com/en/docs/basic_vector_styling.html.

Making a Base Map with an Image

Let's assume you want to use SOS's “earth vegetation” image as your base map. That
image can be accessed from SOS's FTP server:
1. Go to ftp://public.sos.noaa.gov/. Note that this is SOS’s FTP server, where all of
their datasets are available for download, not its website. Go directly to that URL.
2. Click on “land,” then “blue_marble,” then “earth_vegetation,” then click on
2048.jpg.
3. Right-click and select “Save image as.” It doesn't matter where, but make sure
you save it as a .tif image- we'll need it in that format. Save the image in a folder
you can find again. When the dialogue box appears to save your image in
Windows, there will be a field labeled “Save as type.” Click on “All file formats.”
Next, add the extension “.tif” to the name of your image. On a Linux machine,
you’ll have to delete the original file extension first: for example, replace .jpg
with .tif.
6
Windows screenshot:

Linux screenshot:

7
 Next, we need to create a way for QGIS to interpret this image; the software can’t
tell that it’s actually a map until we create something called a “world file.” A world file
is an extra file that QGIS knows to associate with an image file, which gives QGIS a set
of “coordinates” to put the image at. It tells QGIS how big the image is supposed to be,
how to rotate it if at all, and where to place the corners. A world file must have the same
name as the .tif image, but its extension is .tfw, for Tiff World File. This is so that QGIS
knows which image to associate with which world file.

Open up a text editor (such as Notepad for Windows or Text Editor for Linux
machines) and create a blank file. Its name must be the same as your image; only the file
extension should be different. For example, if you have an image named 2048.tif, its
world file should be called 2048.tfw. Also, please note that even though the file
extension is supposed to stand for “TIFF World File”, it’s “.tfw,” not “.twf.” Save it
using the “Save as” feature in the same folder as your image. If the world file is not in
the same folder as its image, QGIS won’t be able to find it.

Now, to create the world file, you have to type in information in a way the
computer expects it to appear. World files have six lines, which each contain one
number:

Line 1: Pixel size in X direction (aka image width)

Line 2: Y-axis rotation
Line 3: X-axis rotation
Line 4: Negative of the pixel size in Y direction (aka image height, with a negative sign)
Line 5: X coordinate of the CENTER of upper-left most pixel
Line 6: Y coordinate of the CENTER of upper-left most pixel

To find this information, we have to know something about our map. First, with
the projection we're in (and with most world maps), there are 360 degrees in the X
direction (the map spans the circumference of Earth) and 180 degrees in the Y direction
(the map only goes from the North to the South Pole, not around the other side). We also
know that our image is 2048x1024 pixels. Knowing this, we can use the following
equations:

Line 1: Pixel size in X direction =

360 degrees / 2048 pixels = 0.17578125 degrees per pixel

Line 2: Y-axis rotation = 0

8
Line 3: X-axis rotation = 0
Line 4: Negative of the pixel size in Y direction =

180 degrees / 1024 pixels = -0.17578125 degrees per pixel

Please note that lines 5 and 6 require the position of the CENTER of the pixel. Since we
already know how many degrees make up a pixel (0.17578125), we divide that number
by two to find the distance from the edge to the center of the pixel.

Line 5: X coordinate of center of upper-left pixel =

-180 degrees + ( 0.17578125 / 2) = -179.9
Line 6: Y coordinate of center of upper-left pixel =
90 degrees - (0.17578125/ 2) = 89.9

Our final text file, 2048.tfw, looks like this:

0.17578125
0.0
0.0
-0.17578125
-179.9
89.9

Type these numbers into your world file, or copy and paste, and save the file. Note
that this world file will only work for an image that is 2048 by 1024 pixels. If your
image has different dimensions, your world file will look different.

Now go back to QGIS and go to Layer -> Add Raster Layer.

9
Select your image (2048.tif), not the world file. Click OK. A box will appear asking you
which CRS (Coordinate Reference System) the data is in. Select WGS 84 and click OK.
Once the image is open, QGIS will know to overlay any other data you might want to
insert in the right place on the map. Your map should now look like this:

10
 Once you have more layers, you can reorder layers by clicking and dragging the
layer names in the layer bar to the left of the viewing window, to make one appear on
top of or below another.

Finding/Manipulating Point Data

For the purpose of this tutorial, I will create a map with the locations of nuclear
reactors plotted against earthquake danger zones.

I’m using a map with a transparent ocean as a base map, so that it’s easy to see
what I’m doing during this tutorial. To insert it, see the “Base Maps with Raster Data”
section of this tutorial on page 6. I found my reactor data in spreadsheet form at
http://www.theguardian.com/news/datablog/2011/mar/18/nuclear-reactors-power-
stations-world-list-map#data. To download it, click on the link right under the title that
says “Get the data.” This takes you further down the website to a link that says “DATA:
Download the full spreadsheet.” Click on that link and save it in a place you’ll
remember.

Once you've downloaded the spreadsheet, you'll notice that while it does include
coordinates, they have both latitude and longitude in the same column. QGIS won't be
able to plot the data unless latitude and longitude are separate. This is frequently the case
with online databases, so I'm going to document the trick I use to separate the column
into two. If you are using a spreadsheet that already has latitude and longitude in
separate columns, skip down a page to the words “Skip to this point if your data does not
require reformatting.”

To split one column into two, follow these instructions:

1) Open a new Google Spreadsheet. (If you prefer using Microsoft Excel, their
tutorial is here: http://office.microsoft.com/en-us/excel-help/split-text-into-different-
cells-HA102809804.aspx. Skip down to the “Skip to this point” message if you do.)
2) Copy and paste your data into the sheet.

11
3) Insert two extra rows next to the coordinates’ column, in the same spreadsheet. Do
this by right-clicking on the column head, circled in green, and selecting “Insert 1 right.”
Do this twice.

4) Click on cell C2. Go to the formula bar located directly above the column headers,
boxed in blue. Type in the formula =SPLIT(B2, “,”). This will split cell B2 where the
comma appears, putting the split halves into cells C2 and D2. It will also make sure the
comma isn't copied into the new cells.
5) Right-click to copy cell C2. Select the remainder of column C, down to where
your data ends. Right-click to paste the formula in C2 into the rest of your selection.

6) Label your new columns “latitude” first and “longitude” second.

Now go to File → Download As → Comma Separated Values (.csv) to save your

spreadsheet.

Skip to this point if your data does not require reformatting, or if you are using Excel
to reformat.

In QGIS, click the “add delimited text layer” button or go to Layer → Add Delimited
Text Layer. Select the file you want to upload (that'll be our nuclear data file).

12
For the X field, pick “Longitude” or whatever column contains longitude coordinates;
for the Y, choose the latitude column. Click OK.
A dialog box will come up prompting you to choose your projection. Pick WGS
84 and hit OK. Huzzah, we have a map of reactor locations! For information about how
to change the color and style of the points, see the next section.

Formatting Point Data

13
 For the earthquake layer of this map, I'm going to want something more than just
locations of earthquakes- I want to somehow add information like earthquake intensity.
The tab delimited variable file I've downloaded from NOAA has all that information; we
just need to show it.

Start by downloading NOAA's significant earthquake database (which contains

“information on destructive earthquakes from 2150 B.C. to the present that meet at least
one of the following criteria: Moderate damage (approximately $1 million or more), 10
or more deaths, Magnitude 7.5 or greater, Modified Mercalli Intensity X or greater, or
the earthquake generated a tsunami”) from
http://www.ngdc.noaa.gov/nndc/struts/form?t=101650&s=1&d=1. Click the link labeled
“Download entire significant earthquake database in tab delimited format.” Once it’s
downloaded, open the significant earthquake database and save it in a place you’ll
remember it.

Import it into QGIS the same way we imported the nuclear reactor dataset, but
when the dialog box that asks you to choose the file comes up, look at the field called
File format. It has three choices: “CSV (comma separated values),” “Custom
delimiters,” and “Regular expression delimiter.” Choose “Custom delimiters.” Check the
tab box if it isn't checked already.

LONGITUDE and LATITUDE should be in the X and Y fields already. Hit Ok. A list of
errors may pop up. These are usually due to a missing set of coordinates. You can ignore
them for now. Another window will pop up prompting you to choose a CRS: pick WGS

14
84 and click OK. Your map should now look something like this:

Now, right click on the layer of earthquake data you just imported and select
“Properties.”

15
 You can format your data to make it show attributes. In this case, I want to show
the intensity of each earthquake via a color scale. Since my data includes a column
labeled INTENSITY, I can tell QGIS to assign a color to each point based on what’s in
that column. However, if you were to try doing so now, you would notice that a bunch of
your data vanishes, because if a point has a null value in INTENSITY, QGIS can’t
assign it a color, so it gets removed. We can fix this by filtering out everything with an
intensity of NULL (which QGIS interprets as having no intensity at all) and making it its
own layer.

Right-click on the layer name of your earthquake data in the layer menu and select
“Open the attribute table.” Click “Select using equation,” which is the button in the top
left corner with an epsilon on it.

In the “Expression” box, type the phrase INTENSITY IS NULL to select the
points with null intensities. Then click Select. Some points will now be highlighted in
the attribute table, and some points on the map will be a different color (yellow instead
of blue, in this example).

16
 Now right click on your earthquake data layer and select “Save selection as.”
You’ll want to click “Browse” and choose where to save your file- you’ll get an error
message if you try to save in the default directory. Check the box that says “Add saved
file to map,” and hit OK. Your selection is now a separate layer that we can format
independently, for example, by assigning it a color value that matches the color bar
we’re using for the rest of the earthquake points.

17
 Now we can reformat our data to give it a color scheme. To do this, go into the
“Style” tab of the Properties window. (Remember that to get to Properties, you right-
click on the layer name and select “Properties.”) First change the bar that says “Single
symbol” to “Graduated.” Next, pick the column you want to be color scaled, which for
me is “INTENSITY.” Pick your color ramp and then click “Classify.” If you want more
than the default five classes, you can adjust the Classes setting found to the right of the
Symbol field. A color scale will appear. Click “Apply.” If you like it, keep it! If not, fool
around some more.

18
Your map now looks like this:

I assigned the null values a light yellow color in the previous step, which isn’t
visible at the moment. Those values don’t show up very well on a white background, so
I’m now switching back to an image with a blue ocean. I’m going to change the color
scheme, and I think I’d like to make the earthquake locations look more blended
together, so that they look more like danger zones than individual locations. So I’m
going to click on the “Change…” button, in the Style tab under Properties. (To get to
Properties, remember that you right-click on the layer name and select “Properties”).
Here you can change the marker transparency, shape, and size. You can click on
“Change…” and select the “simple marker” icon, then set the outline style to “no pen” to
get rid of the outline around the shapes. You can make all sorts of changes here, so I
won’t go into detail. The only thing to remember is that you want your map to be easily
read. For example, you can change the shape of the nuclear power plant layer points to
differentiate them from the earthquake locations. My results are below. (To export this
image, see the “Exporting Maps as Images” section of this tutorial.)

19
Making a map with Regions Colored by Data

So far, this tutorial has only dealt with one type of data: a spreadsheet, which we
use in .csv (comma separated variable) file format. This format is suitable for plotting
data in the form of points or locations, but if you want to show data on a per-region basis
(such as life expectancy per country), the easiest way to do so is to use an ESRI
shapefile. A shapefile is actually a group of several files that must be kept in the same
folder to work. There are a few ways to get them. The first is to simply download a
shapefile from one of any number of websites (a list of suggestions is provided below, in
Option One). The second is to create a new shapefile from a spreadsheet or .csv data,
which is primarily useful if you can’t find a pre-existing shapefile with the right
elements. The easiest way to create your own shapefile is to use CartoDB.com, a
mapmaking website- see Option Two. And finally, you can add new data to a pre-
existing shapefile by adding columns in the layer’s attribute table (see Option Three).

Option One: Getting a Shapefile from the Web

Start by searching for whatever dataset you like as an ESRI shapefile. Some good places
20
to start looking are:
• The Natural Earth datasets: http://www.naturalearthdata.com/downloads/. These
datasets are good as base maps, but probably won’t contain data that can be used
to color-code regions— these maps are usually just borders. However, if you
want a base map composed of vector data, this is a good place to start.
• Geocommons: http://geocommons.com/. User-submitted maps containing all sorts
of data. Search for any kind of map, then click on the "shapefile" button to the
right of the map (if available) to download the shapefiles.
• MapCruzin: http://www.mapcruzin.com/download-free-arcgis-shapefiles.htm. A
luck-of-the-draw eclectic collection of maps and shapefiles. Some shapefiles con-
tain data that can color-code countries, such as historical earthquake occurrences
per country.

For the purposes of this example, I’m going to use the Life Expectancy shapefile
from Atlas of the Biosphere, which is available at
http://www.sage.wisc.edu/atlas/maps.php?datasetid=12&includerelatedlinks=1&dataset
=12. Although this dataset is labeled “Life Expectancy,” it actually has numerous other
statistics included in it, such as infant mortality rates, access to safe water per country,
etc. You can click on the link that says “Download a GIS grid of this data (ESRI ArcGIS
format). This will open up a new page, with a button that says “Download Now.” Click
on it.

It will download a zipped file. At the bottom of your screen, a Downloads bar should
21
appear with an icon that says “lifeexpectancy.zip.” We need to unzip it, so double click
the icon to open the zipped folder. Once the folder opens, you should see all the files
inside it. Go up a folder in your file tree so you can see the folder the files are stored in,
lifeexpectancy.zip. Right-click it and select “Extract all.” Make sure you extract the
files into a location where you’ll be able to find them again.

Now, go into QGIS and start a new project by going to Project -> New. Click the
“Add vector layer” button. You can also click on the Layers menu at the top of the page
and select Add Vector Layer.

Click the Browse button next to the Dataset field and navigate to your extracted files.
You want to upload the one that has the .shp extension— that’s the actual shapefile.
Click Open.

A window will appear asking you to specify the CRS— that’s the Coordinate Reference
System. SOS uses WGS 84, also known as EPSG:4326. Select this if it isn’t already
selected and click OK.

22
Your map should now look like this:

To style this shapefile, you’ll want to choose a column of data such as Life
23
Expectancy and tell QGIS to color countries based on the values in that column. In the
interest of not reinventing the wheel, please see Ujaval Gandhi’s tutorial on the subject,
which may be found here: www.qgistutorials.com/en/docs/basic_vector_styling.html.
Some points to be aware of:
• Assuming you’ve followed the steps above, you can skip to step 4 of Ujaval’s tu-
torial. Steps 1-3 are instructions on how to get the data into QGIS.
• When you color countries according to values in a column, countries with no val-
ues in that column will get grouped with the countries that have the smallest val-
ues. QGIS assigns those countries the value “-99” for some reason.
• If you only get two colors in your map after assigning it a color scheme, try
changing the mode or adding more classes.

Once you have styled your map to your satisfaction, please skip to the “Exporting
Maps as Images” section of this tutorial. You’ll also add legends, colorbars, and other
map features in that step.

Example of Styled Map: Access to Contraception per Country

I used the graduated symbols option with seven equal-interval classes and the
“Blues” color ramp to make this map.

24
Option Two: Creating a Shapefile Using CartoDB

CartoDB is a website that allows users to upload spreadsheets of geographic

information and plot them on a map of the world. Unfortunately, the map CartoDB uses
is in the wrong projection to work with SOS (see the write-up on CartoDB for more
information), but it’s still a useful tool for converting CSV files to shapefiles. The
process of converting text or image information to information that is associated with
location coordinates is called georeferencing. CartoDB can do that for us.

Getting CartoDB

In this case, getting the mapmaking software is very simple: go to

www.cartodb.com and create an account. CartoDB designs plans based on storage space.
If you’re importing a lot of data for individual maps, or creating a lot of maps, you’ll
need to get one of the paid versions. If not, scroll down past the descriptions of the paid
versions and click on “free version.”

Importing the Data

For this tutorial I will be using the International Telecommunications Union’s

database of landlines and mobile phones registered throughout the world. These files
weren’t quite in the format CartoDB can use, so I had to mess with them in Excel first.
This is frequently the case with databases on the web, and since this particular
spreadsheet’s issues were very specific, I won’t go into how I solved them. If you would
like to recreate this dataset, please see the collection of datasets provided along with this
tutorial. The file is called Fixed_tel_2000-2012.csv.
A note about CartoDB’s preferred formats: Please see the developer help website at
http://developers.cartodb.com/documentation/using-cartodb.html for specifics on
formatting.

25
 To start, go to the green Dashboard button in the top-right corner of the page. This
will get you to your dashboard page, where your datasets are stored in table format.
Click on the large “+” button labeled New Table.

Select “Select a file,” and find the spreadsheet you need. CartoDB can take Excel,
CSV, TSV, ESRI Shapefiles, KMLs and KMZs, GeoJSON, GPS eXchange (GPX), OSM
and BZ2, OpenDocument Spreadsheets (ODS), and SQL. See
http://developers.cartodb.com/documentation/using-cartodb.html for more information.
Click Open. Your table will appear.

At the top of the screen, just above your table in the left hand corner, will be the
words “Table view” and “Map view.” Click on “Map view.” This window will appear. If
it does not, click on the Options button in the top right corner of the screen and select the
“Georeference” option.

26
 Table View and Map View

Click “Georeference.”

On the next window that appears, you’ll want to click “You have a column
identifying administrative regions” since our data is referenced by country, not latitude
and longitude coordinates. In the “In your table the polygons are identified by…” field,
select “field_1.” This means that the countries are listed in the field_1 column in your
dataset. If you look at the column heading in CartoDB’s Table View over the column
with country names, you’ll see that CartoDB has labeled it “field_1.” You want your
data in “the World” by “Countries” just like the default settings say, so click
“Georeference.”

27
 A screen will come up with two boxes; one will be greyed out and say “No point
data available for your selection” and the other will say “Georeference your data with
administrative regions.” It should be highlighted. Click “Continue.” The data will take a
minute to render.

28
 CartoDB will then give you a message “X out of Y rows were successfully turned
into polygons!” If the number of rows turned into polygons was lower than you
expected, go through your data in Table View to make sure country names are spelled
correctly. If you have rows with no value in field_1 (that’s the column with the country
names), CartoDB thinks it’s misreading the rows and will count them along with the
rows it couldn’t transform, so check to see if that’s the cause of the discrepancy. If it
isn’t, go into “Map view” and find the countries that aren’t overlaid with a color— these
are the ones CartoDB couldn’t parse. For example, Iran in this dataset was labeled as
“Iran (I.R.).” CartoDB couldn’t recognize that so I changed it to “Iran” by double-
clicking the name to edit the text.

You may have to delete and re-upload your table to make the changes take effect,
in which case you should make your changes in the original spreadsheet, save that
spreadsheet as a .csv file (or whatever format was originally used), and upload it just as
you did before.

Now click on the Options button in the top-right corner and select Export.

Select “shp” as your file type and save your file in a place you can remember it. Note
that shapefiles come with several different documents, all with the same name, but with
different extensions, such as filename.shp, filename.prj, and so on. If you decide to
move the shapefile after it’s saved, you’ll need to move all of the files with it.

29
 The shapefile will be downloaded onto your computer as a zipped file. Unzip it by
right-clicking the file and selecting “Extract all.” Extract the files into a place where
you’ll be able to find them again.

Now, go into QGIS and click the “Add vector layer” button. You can also click on
the Layers menu at the top of the page and select Add Vector Layer.

30
Click the Browse button next to the Dataset field and navigate to your extracted files.
You want to upload the one that has the .shp extension— that’s the actual shapefile.
Click OK.

A window will appear asking you to specify the CRS— that’s the Coordinate
Reference System. SOS uses WGS 84, also known as EPSG:4326. Select this if it isn’t
already selected and click OK.

Your map should now look something like this:

31
 In the interest of not reinventing the wheel, to learn how to style vector layers,
please see this tutorial by Ujaval Gandhi:
www.qgistutorials.com/en/docs/basic_vector_styling.html. Some points to be aware of:
• Assuming you’ve followed the steps above, you can skip to step 4 of the tutorial.
Steps 1-3 are instructions on how to get the data into QGIS.
• When you color countries according to values in a column, countries with no val-
ues in that column will get grouped with the countries that have the smallest val-
ues. QGIS assigns those countries the value “-99” for some reason.
• If you only get two colors in your map after assigning it a color scheme, try
changing the mode or adding more classes.

When you have styled the data to your satisfaction, please see the “Exporting
Maps as Images” section for the next steps.

32
Example of Styled Map: Landlines per Country

For this map, I used the Graduated Symbols style with five equal-interval classes
and the “Greens” color ramp.

Option Three: Joining a File to an Existing Shapefile

To “join” a shapefile and a spreadsheet means to add the data in the spreadsheet to
the data in the shapefile. For example, if you have a shapefile of countries and a
spreadsheet of population data for the same countries, to join the file, you would tell
QGIS that the two columns with country names in them should be matched. Then, in
addition to the shapefile’s original data for each country, the shapefile will contain the
population data as well. For this example, I will be using the Life Expectancy shapefile
from Atlas of the Biosphere, which is available at
http://www.sage.wisc.edu/atlas/maps.php?datasetid=12&includerelatedlinks=1&dataset
=12. This is the same dataset used in the Option One section of this tutorial. For the
spreadsheet, I will be using the Language Spreadsheet provided by Brown University. It
describes the most prevalent language of countries in the 1500s. It can be found at this
link: http://www.econ.brown.edu/fac/Louis_Putterman/Language%20Spreadsheet.xls.
To use it as described in this tutorial, simply save it on your computer, then open it. Go
33
to File -> Save as. Save the spreadsheet as a .csv file by changing the file extension from
.xls to .csv.

First, you need to open the spreadsheet in QGIS. To do so, go to Layer -> Add
Vector Layer and select the CSV file you just created. Note that you’re not adding it as a
delimited text file, which is what we have done in the past. Click Open. It will appear in
your Layers bar on the left side of your screen, but no data will appear on your map,
since QGIS doesn’t know how to interpret it yet.
Please follow the steps in Option One for importing a shapefile to import the Atlas of the
Biosphere shapefile. Once you have your shapefile, right-click on it in the Layers menu
on the left side of the screen and select “Open attribute table.”

You need to find an identifying column of shapefile data that will match up with a
column on your spreadsheet- in our case, we’re looking for the country codes. We’re
using those instead of names because QGIS has to see exactly the same word in the
spreadsheet as it sees in the shapefile’s data, or it won’t be able to match the two.
Differences in abbreviation or spelling errors are easier to avoid if you’re using three
letter codes instead of names, since the codes are standardized. Look through the
attribute tables of your shapefile and your speadsheet to find the name of the column
that contains the country codes. As it turns out, those are under the column labeled
“Code” on the spreadsheet and the column labeled “WB_CNTRY” in the shapefile’s
attribute table. Once you’ve found both, close both attribute tables.

34
 Now open the shapefile’s properties window. You can do this by right-clicking the
name of the layer in the layer bar to the left and selecting “Properties” or by simply
double clicking the layer name.

Go to the “Joins” tab and click on the “Add” button- it looks like a green plus.

A window labeled “Add vector join” will appear. Select the layer you want to join
to the shapefile (our Language Spreadsheet layer), the join field (that’s the column of the
spreadsheet, “Code”) and the target field (that’s the column of the shapefile’s attribute
table, “WB_CNTRY”). Make sure that “Cache join layer in virtual memory” is checked.
Click OK on the “Add vector join” window and the “Properties” window.

35
 Your shapefile should now have added the information from the spreadsheet to its
attribute table. You can check this by simply opening the attribute table by right clicking
on the layer name and selecting “Open attribute table,” then finding the new columns in
the table.

To color the countries according to the data you just added, please see Ujaval
Gandhi’s tutorial on the subject:
www.qgistutorials.com/en/docs/basic_vector_styling.html. Some points to be aware of:
• Assuming you’ve followed the steps above, you can skip to step 4 of the tutorial.
Steps 1-3 are instructions on how to get the data into QGIS.
• When you color countries according to values in a column, countries with no val-
ues in that column will get grouped with the countries that have the smallest val-
ues. QGIS assigns those countries the value “-99” for some reason.
• If you only get two colors in your map after assigning it a color scheme, try
changing the mode or adding more classes.

When you have styled the data to your satisfaction, please see the “Exporting Maps
as Images” section of this tutorial.

36
Example of Styled Map: Primary Languages of the World’s Countries in the 1500s

For this map, I used the Categorized style based on the column labeled “Language
Spreadsheet_Language.” I used the “random colors” color ramp to delineate each
language.

Exporting Maps as Images

To export a map as an image, you use QGIS’s print composer. This is also where
you can add titles, compasses, and scales, but due to the nature of our maps, we usually
don’t. It’s pretty obvious which way north is on a globe. SOS recommends that you keep
any text that must be stationary on the map, like labels, close to the equator, so that it
isn’t warped when the map is transformed onto the sphere. However, the preferred way
to put legends, images, and labels onto the sphere is to make a PIP (PIP stands for
‘picture in a picture’). For a detailed description of PIPs, please see SOS’s Content
Creation guide, found at http://sos.noaa.gov/Docs/SOSContentGuidelines.pdf. For a
description of how to create a PIP, see the “Map Legends and Color Bars” section of this
tutorial.

37
 Right-click on your image layer and select “Zoom to layer extent.” Then go to
Project -> New Print Composer. Name it anything you like.

Click on the “Add Map” button and draw a rectangle that fills the whole white
workspace by dragging your mouse across the workspace. Try to make this as exact as
you can. Whatever is on QGIS’s main screen will appear in the box. Note that if you
want to change what’s appearing in your map box, you can change it in the main QGIS
window, and then click the “Update” button under Item Properties. Whatever you most
recently clicked on in the composer will get updated. For example, if you have a legend
and a map arranged in the composer, only the one that you last clicked on will get
updated. Now look at the menus on the right. First, next to Presets under the
Composition Tab will be a box with a preset paper size in it. Click on it and select
“Custom.” Then change the width and height to 6.84 inches and 3.41 inches,
respectively. This is because we want our export resolution to be 300 dpi (dots per inch,
or pixels per inch) and we want the final resolution of our map to be 2048x1024. 2048
38
divided by 300 is approximately 6.84, and 1024 divided by 300 is 3.41. You’ll probably
have to change the units to inches instead of mm. Make sure your export resolution
really is at 300 dpi.

“Add Map” button

Click the “Item properties” tab, then scroll down to “Extents.” To make your map

Export as Image Button

39
fill the screen as much as possible, change the X values to -180 and 180, and the Y
values to -90 and 90. You should see the map fill the entire white space. Check the Scale
value under Item Properties. For a map sized 6.84 by 3.41 inches, resolution 300 dpi, the
scale should be 114891360. If this is incorrect, it causes problems with the export. Then
click the “Export as Image” button. SOS prefers JPEG or PNG, but can accept most
common image formats. Congratulations, you have an SOS-ready map!

As a quick reminder, remember that trying to export very high-resolution images

frequently does not work. If you are experiencing odd artifacts in your exported image,
try switching out your raster layer for a lower resolution one. I did not experience any
problems with 2048.tif, but I did try to create a map with four times the resolution and
got all sorts of issues.

Map Legends and Color Bars

SOS recommends that users create labels, legends, and text in the form of “PIPs,”
(picture in a picture) which are images that can be projected on the sphere. This makes it
possible to avoid the warping that would accompany text placed directly on a map, since
the closer to the poles the text is, the more it is warped when it is projected onto the
sphere. QGIS allows you to create PIPs using the print composer. This tutorial
(http://maps.cga.harvard.edu/qgis/wkshop/export_map_print.php) is a good resource for
learning to create most common map accessories. Once you have a legend that you like,
you should save it as a separate image. You can accomplish this by unchecking all of
your map layers so that none of them are visible in the print composer (note that you
have to have them visible when you open the composer, or the legend won’t be created
in the first place), and then exporting the image as a PNG. You may want to crop it down
in an image editor such as GIMP or a Microsoft Office product like Word or
PowerPoint.

For example: if I wanted to create a legend for my earthquake data, I would first open
the print composer with one or more earthquake layers visible and create a legend as
described in the tutorial that was linked to above. Then I would uncheck all the layers
and click “Update” under the “Item Properties” tab to ensure that the legend is the only
thing visible in the print composer, and click “Export as Image.”

Exporting Maps as Layers

40
 Sometimes it can be advantageous to have all the layers of your map as separate
image files, so that they can be turned off and on as you wish. For example, I’m saving
the nuclear power plant locations as a separate PNG from the earthquake data, in case
some users just want to talk about one or the other. To save the layers of your map as
separate files, simply uncheck the boxes next to each of the layers that you don’t want
visible, and arrange the remaining layer(s) for the print composer just as you would for a
full map (For guidance on how to do this, please see the section labeled “Exporting
Maps as Images.”)

I also want to create layers of different years of earthquake data, because I have
data from four thousand years ago and it would be neat to show the earthquakes
accumulate, layer by layer. I’m going to split it up into chunks of 200 years. To do so,
I’m going to create a layer of data for each one, and then I’m going to arrange each layer
in the print composer.

To split up the layers of data, right-click on the layer name and select “Open
attribute table.” Click on “Select features using an expression.” If you don’t remember
from the steps above, that’s the button at the top that has an epsilon on it. In the
Expression window, we need to tell QGIS to select all the elements of column YEAR
that have values within a certain range. The formula to do so is ‘ “YEAR” <= -2000.’
Type this into the Expression window. Do not include the outer set of quotation marks
(the single quotes) or the period after -2000. To explain what this formula means: We
want to select the values in the YEAR column, by putting the column name in quotes
(without the quotes, QGIS thinks you’re referring to the column itself, rather than each
value within the column). The first range of years we’re selecting is anything older than
2000 BCE, so we use the formula “YEAR” <= -2000 to say “The values in column
YEAR that are less than (<) or equal to(=) -2000.” Note that although some operator
buttons are provided above the Expression window, you can use the keyboard to type
characters that aren’t provided there. Click on Select.
Once you have a selection, right-click on the layer name and click “Save selection
as.” Save as an ESRI shapefile in a folder you’ll remember and import the resulting
layer to QGIS by checking the “Add saved file to map” and pressing OK.

Next, we need to select values between two numbers. We do so by adding the

keyword AND to tell QGIS to find points that satisfies both conditions. To find
earthquakes between the years 2000 and 1800 BCE, not including 2000, the expression
is “YEAR” > -2000 AND “YEAR” <= -1800. Note that “YEAR” > -2000 AND <= -
1800 is not a valid statement.

41
 Repeat these steps as many times as you need. For this map, since our color style
drops any point with an intensity value of NULL, I’m also making a layer with the
command “INTENSITY” IS NULL added onto my year specifications for each set of
200 years, like so: “YEAR” > -2000 AND “YEAR” <= -1800 AND “INTENSITY” IS
NULL.

Since it would be a pain to go through and redo the color style for each layer, I’m
going to go to Properties (right-click layer name and select Properties) for my
earthquakeData layer and save the style I’m using as a QGIS layer style file. I’ll do the
same on the null intensity layer. As a result, I can use the “load style” button to copy
those styles to any other layer. However, this method seems to be a bit buggy, so
another, simpler way to do this is to right- click a layer and select Copy Style, then right-
click the new layer and select Paste Style.

42
 Once all the layers are created, it’s time to export them all as PNGs. Remember to
save them with useful names! See the Content Creation guide on SOS’s website for
nomenclature guidelines. Arrange your images in the Print Composer the same way we
did for a full map, but this time, make sure that only the layers you want to export are
visible in QGIS’s main window. Make sure that you right-click on one of the layers that
covers the full extent of the map, like the background image layer, and select “Zoom to
layer extent,” or your points will be in the wrong places. You can use one composer to
export all your images; just select the layers that you want in each one and press the
“Update preview” button on the Item Properties menu. See the “Exporting Maps as
Images” section of this tutorial for detailed instructions.

43
 Remember to make sure that your DPI is set to 300, your paper size is custom and
set to 6.83 by 3.41 (this gives us the correct pixel size of approximately 2048 by 1024),
and your Extents are set to -180, -90, 180, and 90. See the Exporting Maps as Images
section for details.

Note that QGIS does not export these maps with transparent backgrounds. You
will have to remove those afterwards in a program such as Photoshop, Gimp, or
ImageMagick. I’ll go over that in more detail below. The thing to think about is that
you’ll want the background of your images to contrast strongly with your data points.
You may find it necessary to insert a background image of a solid color (and a world
file, to make sure the image is in the right place) to contrast with your points. Be aware
that if your points are slightly transparent, as ours currently are, adding a background
color will change the color of the points. Because of this limitation, we aren’t adding a
background color in this tutorial.

Making Image Backgrounds Transparent

To make the map layers’ backgrounds transparent, you need image editing
software. This tutorial will cover how to do so with GIMP, which is freeware and can be
used on either Windows or Linux. If you have access to Adobe Photoshop, that works
too.
44
 To download GIMP, go to http://www.gimp.org/downloads/. Download by
clicking on the link after “Gimp 2.8 is now available at.” It will get you to a page full of
download links. The top link should be the one you need- it will have the word LATEST
in the URL. Follow the instruction in the setup wizard. When you open it (you can do so
by going to the start menu and searching for it) it won’t be full screen. If you want to set
the transparency of one image, this is all you need, but if you want to batch process a
bunch of images (we do), you’ll need to install the plugin BIMP as well. To do so, go to
http://www.alessandrofrancesconi.it/projects/bimp/ and click the BLUE “Download
here” button. DO NOT CLICK THE GREEN DOWNLOAD BUTTON!! It is an ad that
will download malware onto your computer. Once you’ve downloaded the file, you’ll
end up with a zip file in your Downloads folder, which you’ll need to extract into
GIMP’s plugins folder. To do so, right-click the BIMP zip file, select “Extract all,” and
in the box that says “Files will be extracted into this location,” browse until you find
Gimp’s plugins folder. In Windows, that should be found at the file path
C:\Users\your_user_name_here\.gimp-2.8/plugins. So to find it, go to the folder
Computer, then Disk C: (your local hard drive), the folder Users, then the folder with the
same name as your username, then to the folder labeled gimp 2.8, then the folder called
plugins. Select the plugins folder and click OK. Once you’ve done that, open the plugins
folder by going to the Start menu, clicking on Computer, and following the file path
above. Once you’re in the plugins folder, go to bin -> win32 and copy everything in
win32. Paste it back into the main plugins folder. Your plugins folder should now look
like this:

45
 Now, open GIMP. When you click File, you should see the option Batch Image
Manipulation. Click on it.

In the Manipulation Set box, click the “Add” icon to add a new transformation. A
menu will appear; select Other Gimp Procedure.

46
Go down to plug-in-colortoalpha. Click on “color to remove.”

Change the color value to #FFFFFF (this is the hexadecimal code for the color
white). Click ok on the color picker window and the manipulation set window.

47
 In the original Batch Image Manipulation Plugin window, click “Add images.”
This is not the same button as the “Add” button you just used to pick your Select the
images you want to manipulate and click “Add.” WARNING: BIMP OVERWRITES
THE ORIGINAL FILE IF YOU SAVE THEM IN THE SAME FOLDER!! If you don’t
want to overwrite your originals, select a different Output Folder than the ones the
images came from.

Click Apply. For some reason, I had to click Apply three times before my images
actually became transparent, and they only did so when saved to a new folder. I’m not
sure what was going on there. But that should be good— you now have your layers of
earthquake data, in transparent PNG format!

Connecting Points of Data with Lines

Sometimes it’s useful to be able to connect the dots of a dataset, such as in the
case of animal locations. QGIS doesn’t provide this functionality on its own; we have to
download a plugin to do it for us. For this example, we are going to use the leatherback
48
turtle track information provided by www.topp.org, the Tagging of Pelagic Predators
project’s website. The turtle track may be downloaded here:
www.gtopp.org/data/downloads.html or from the datasets provided along with this
tutorial. The file name is “leatherback 41708.txt.”
Open QGIS. At the top of the window, go to Plugins -> Manage and Install Plugins, then
search for “Points to Path” and click Install. Please note that if you search for “Point
to...” instead of “Points to…” the plugin will not appear in the menu.

Run the plugin afterwards by clicking on the Plugins menu again; it will have
appeared in that menu. Open the plugin.

You will see a window pop up that looks like this:

49
 Currently, it’s on the “General” tab. There are also a “Date Format Reference” tab,
which will be useful later, and an “About” tab. On the “General” tab, there are six
fields:
1. Input point layer: The layer of points that you want to connect. In our case, this is
the turtle tracking layer.
2. Point group field: The plugin is looking for a column that has the same charac-
ter(s) for all the points that will be part of one path. The software just needs to
know which points it should add to the line, so it asks for a column with a com-
mon identifier. For example, if I have a spreadsheet that contains location infor-
mation for one leatherback turtle, I would need an identifying column that has the
number 1(or any other character) in it for every row of data. If I had a spreadsheet
with information for two turtles, turtle 1 would have the number 1in the identifi-

50
cation column, and turtle 2 would have the number 2, so that Points to Paths
knows to make two paths for the data.
Example for one leatherback turtle:

Example for two leatherback turtles:

51
3. The point order field is the field that tells the plugin which order to connect the
points in. What the plugin wants you to do is specify a column that has the date
the location measurements were taken. For our data, and possibly in other spread-
sheets, you may run across a column called “date.sec” that doesn’t look like a
recognizable date format. A computer will be able to use that information without
being told what format it’s in, so if you pick the date.sec column for the point or-
der field, leave the date format field blank. However, most of the time, you’ll
have a more traditional date column. In this case, you’ll need to tell the plugin
what format it’s in in the following column.
4. The date format field’s purpose is to tell the plugins which format your date in-
formation is in. For example, say you have a date column in the format
MM/DD/YYYY. (To determine the format of your spreadsheet’s date column,
right-click on the layer name and select “Open attribute table.” You’ll be able to
see the column that has your date info.) The plugin needs to know that the first
number it reads is the month, the second is the day, and the third is the year. It al-
so needs to know that a number ends when it is followed by a “/” symbol. So we
type in the symbols “%m/%d/%Y,” which basically translates our date into a for-
mat the plugin can read. “%m” means month, “%d” means day, etc. See the “Date
Format Reference” tab at the top of the plugin’s window, it will give you the
codes needed to identify each type of date information.

52
 Date Format Reference tab

Date column in
attribute table. Give
the plugin this format.

5. Gap period: You don’t need to worry about this field for this dataset
6. Output Shapefile: Under the “Output shapefile” field, select a name for the animal
track you’re creating. You can use “Browse” to pick a folder to save it in. The an-
imal track will be saved as a shapefile.

After you click OK on the plugin window, a window will pop up asking you if you
want to import the new shapefile. If you click “Yes” and the shapefile doesn’t appear,
simply import it as a vector layer manually. (See the “Base Maps with Vector Data”
section of this tutorial.) It will be saved in your top directory, unless you specified a file
path while naming your output shapefile. A line that goes through your points will
appear on your map.

53