Creating radial flow maps with ArcGIS

By Mamata Akella, Esri Design Cartographer

Flow Map blog thumbnail

Flow maps show the movement of some phenomenon, normally goods or people, from one place to another. Lines are used to symbolize the flow, typically varied in width to represent differences in the quantity of the flow. In broad terms there are three main types of flow map: radial, network and distributive. Radial flow maps have a spoke-like pattern because the features and places are mapped in nodal form with one place being a common origin or destination. Network flow maps are used to show interconnectivity between places and are usually based on transportation or communication linkages. Distributive flow maps typically show the distribution of commodities or some other flow that diffuses from origins to multiple destinations.

In this blog, I show you how to create a radial flow map that depicts multiple origins converging on a single destination. For these types of flow maps, the precise route of flow is not as important as the general direction. A good example of this type of map is the airline route map you might see in the back of an airline magazine with multiple origins travelling to an airline hub.

The map that we created to demonstrate this concept was made for this year’s Making Beautiful Maps technical workshop at the Esri International User Conference. Entitled ArcAttendees: The 2011 Esri UC Global Gathering, the map shows what cities people came from to attend the User Conference (figure 1).

TPFS figure one

Figure 1: A radial flow map showing the city of origin for attendees to Esri UC 2011 in San Diego, CA.

This blog covers the following topics to show how this map was made:

  • Data requirements
  • Flowline construction
  • Data processing steps
  • Symbolization
  • Projection and layout

Data requirements

Esri’s Marketing Department supplied a table of attendee data (figure 2), that contained the following information:

  • a field GEONAMEID that is a unique identifier connected to our cities dataset;
  • the city and country that attendees were coming from (FIRST_ENG, FIRST_CNTR,);
  • the xy coordinates for each city of origin (POINT_X, POINT_Y); and
  • the xy coordinates for the destination city, San Diego, CA (DESTINATION_X, DESTINATION_Y).

TPFS figure one

Figure 2: 2011 Esri UC attendee data

We also used a feature class of cities also attributed with the same GEONAMEID field that exists in the attendee data; and a continent feature class used to symbolize the flow lines according to continent of origin. The continent feature class was also used to symbolize the background on the final map.

Flowline construction

To create the flowlines, we used the new XY to Line tool in ArcGIS 10. The XY to Line tool uses origin and destination values as inputs from a table and creates a new linear feature class that represents the path connecting these two points.

The XY to Line tool is located in the Data Management Toolbox > Features Toolset. Figure 3 illustrates the parameters we defined for the tool:

TPFS figure one

Figure 3: XY to Line Tool

  • Input Table: Attendee data (in .csv format)
  • Output Feature Class: feature class named UCAttendee_XYtoLine
  • Start X Field: POINT_X field from the attendee data
  • Start Y Field: POINT_Y field from the attendee data
  • End X Field: DESTINATION_X field from the attendee data
  • End Y Field: DESTINATION_Y field from the attendee data
  • Line Type: here there are four options: GEODESIC, GREAT_CIRCLE, RHUMB_LINE, or NORMAL_SECTION. For this map we needed to create curved lines that represent the shortest great circle path between the origins and destinations. Geodesic lines create a straight line based on a spheroid. Great Circle lines are similar but are used for spheres. A rhumb line calculates the line of constant bearing and normal section creates a straight line directly on a spheroid. For this application, any of the first three of the options would work to approximate the arc of a great circle between the origin and destinations. We used GEODESIC lines simply because it gave us the best appearance when projected.
  • ID (Optional): here we added the GEONAMEID field from our data to give each flowline created the same ID as the input table which is important for later steps
  • Spatial Reference (Optional): we kept the default GCS_WGS_1984 (we’ll describe later in the blog how we modified the projections for our maps).

Figure 4 shows the results of running the XY to line tool on the attendee data displayed using the default GCS_WGS_1984 coordinate system and figure 5 shows the associated attribute data for the new linear feature class. There were 1532 unique origins in the attendee data with a line created between each of those and the destination in San Diego.
TPFS figure one

Figure 4: Result of the XY to Line Tool

TPFS figure one

Figure 5: The attribute table for the results of the XY to Line Tool

So we now have our basic lines and can begin to attribute, project and symbolize them to show the data effectively.

Data Processing and Symbolization

Since we wanted to symbolize the unique lines by continent of origin and also add some points to the map to indicate the cities of origin, we joined the city dataset attributes to the flowline dataset attribute table based on the GEONAMEID field. This adds the continent field data to the flowline dataset giving us a way to differentiate each flowline qualitatively on the final map.

Symbolizing the Flowlines

We symbolized the flowline features using Categories > Unique Values (figure 6) and the Continent field we joined to the flowline data. We chose highly saturated colors that contrasted well with the dark blue background being used to symbolize the map background (figure 7).

TPFS figure one

Figure 6: Unique colors assigned to the lines by continent

TPFS figure one

Figure 7: Vivid colors used to symbolize the flowlines contrasted well with the dark blue background map detail

Symbolizing the City Points

We symbolized the city points using a circle from the Esri Default Marker font. We set the size to 3 points and made them a bright yellow (figure 8). The use of a highly saturated yellow was chosen to compliment the flowlines and create a fiber optic appearance.

TPFS figure one

Figure 8: The Symbol Property Editor dialog for the city point symbology

Projection and layout

The way in which we arrange the data on the final map needed careful consideration given the distribution of the flowlines. Figure 9 illustrates the data focused on the USA and shows clearly that the lines coalesce to the extent the effect is seriously compromised. We needed to find a way of disentangling some of this detail.

TPFS figure one

Figure 9: The majority of attendees are from the contiguous 48 United States

The majority of attendees came from the contiguous 48 United States. As figure 9 illustrates, this data obscures a lot of other detail and, in particular the pattern in relation to international attendees. Here, then, is a perfect example of the need for using two maps to show the different distributions: an inset map for attendees from outside the contiguous 48 and a main map of those from the contiguous 48. The decision was taken to illustrate the data this way due to the higher number of attendees from the contiguous 48 states while ensuring prominence is given to international attendees on a separate map.

The world inset map is projected using the Winkel tripel projection that allowed the lines to curve in a visually appealing way. It’s a good compromise projection where shape, area, distance and direction aren’t highly distorted (figure 10). We modified the central meridian to -120 degrees to center the map on San Diego, CA.

TPFS figure one

Figure 10: Defining a World Winkel tripel projection for the world inset map

For the map of the contiguous 48 states, we used the World Vertical Perspective projection to create the effect we often see in airline route maps of flowlines moving towards a hub. Here, we wanted San Diego on the horizon with lines travelling towards the city. The fact San Diego is in the lower left corner of the USA meant we were able to modify the projection to achieve this effect as all lines were originating between north and east of the destination. We adjusted the Longitude of Center to -90 degrees (near the tip of the Yucatan Peninsula) and the Latitude of Center to 25 degrees (around southern Louisiana). Finally, we also modified the Height value to give the finished effect (figure 11).

TPFS figure one

Figure 11: The Projected Coordinate System Properties for the contiguous 48 map

The final flowline map is a combination of using new tools in ArcGIS, an interesting dataset and a focused design to create a visually stimulating and effective representation. With the addition of some tabulated statistics summarizing travel information by continent for the Esri UC attendees and arranging the saturated map foregrounds against a light gray background and white map surround we created a strong figure-ground. The use of the new XY to Line tool provides a powerful way to begin to visualize data in innovative ways using ArcGIS, in particular to begin to explore the use of flow maps to depict the characteristics of related points.

Future blogs will explore ways of using ArcGIS to create distributive and network maps where our focus turns to displaying different flowline quantities either between an origin and destination of through a transportation network.

This entry was posted in Mapping and tagged , , , , . Bookmark the permalink.

Leave a Reply

12 Comments

  1. Pingback: Using GIS to Showcase College Acceptance | GIS Lounge

  2. isull says:

    What other resources does ESRI have for radial flow maps? Have there been any follow-on blog postings/forum discussions about creating distributive and network maps? Have there been any additional discussions about displaying different flowline quantities between (1) orgin & destinations and/or (2) through a tranportation network?

  3. daltongis says:

    I was struggling to find where the Winkel tripel projection was available in the ArcGIS coordinate systems list. For anyone else who may want to use this projection, it is located under Coordinate Systems->Projected->World or World-Sphere->Winkel.

  4. nawrotzki says:

    Thanks for the great blog!
    I have followed the described steps to generate a map of global migration flows (between major regions/continents). However, I have been running into an issue that I seem to be unable to solve. The problem is that the lines that I generate with the XYToLine tool in ArcGIS always connect two points (centroids of the region/continents of origin and destination) on the map by the shortest possible line and thus sometimes go outside the map (e.g., connecting Latin America with South-East Asia). Unfortunately, the Winkel Tripel projection does not solve this problem entirely. I have also tried to use other projections such as Fuller or Cube but using these projections the lines are still interrupted and do not show the nice bend as in the Winkel projection. Are there any suggestions on how to solve this issue? Thanks a lot!

    • abuckley says:

      As you note, the liens always represent the shortest path between the origins and destinations. this, if that shortest path is a straight line given the locations of the origins and destinations and the map projection you are using, the lines will be straight. If you want to have more cartographically pleasing lines to deal with something like the projection you are using or the fact that they happen to be straight and you want them to curve, you could start an edit session and reposition the lines any way you want. This would be a final cartographic touch to add to a limited number of lines since you probably wouldn’t want to do this for all the lines if there are many of them!

  5. nlemay says:

    Dear abuckley, I was at the ESRI 2013 UC in San Diego and I was wondering if it would be possible to access the resources from your talk?

  6. carlynne says:

    Thanks a lot for this useful blog!

    I am trying to make web map that shows where in the world refugees from the nine camps along the Thailand-Burma border have resettled. I envision my map similar to the one shown here; however, the origin will depict 9 camps with lines shooting out to multiple cities across the world with about 75% traveling to US cities.

    For a flow web map of this type, can I use the World Winkel tripel projection or must I use the web mercator because the data will be overlaid on base layers that use web mercator, such as grey basemap.

    Thanks for any feedback!
    Lyndy

    • abuckley says:

      It doesn’t mater which projection you use when you create the lines – they will always represent the shortest distance ON THE EARTH’S SURFACE between the origin and destination points. If you project the lines after you create them, the shape will vary but it will still represent that line of shortest distance. So your choice of projection really relates to how you want your final map to look. Your choice of projection should reflect 1) the projection that best illustrates your data (for example, using web Mercator is not a great choice for something that is happening at the poles), and 2) if you want your data to be mashed up on other peoples’ maps (if so, web Mercator is the most commonly used projection so you want to provide your map in something that others will most often use.) Hope this helps!