Archive

QGIS

Today’s post was motivated by a question following up on my recent post “Details of good flow maps“: How to create arrows with gradients from transparent to opaque?

gradient_arrow

The key idea is to use a gradient fill to color the arrows:

gradient_arrow_settings

It all seems perfectly straightforward: determine the direction of the line and set the gradient rotation according to the line direction.

But wait! That doesn’t work!

The issue is that all default angle functions available in expressions return clockwise angles but the gradient rotation has to be set in counter-clockwise angles. So we need this expression:

360-angle_at_vertex($geometry,1)

Happy QGISing!

Advertisements

In my previous posts, I discussed classical flow maps that use arrows of different width to encode flows between regions. This post presents an alternative take on visualizing flows, without any arrows. This style is inspired by Go with the Flow by Robert Radburn and Visualisation of origins, destinations and flows with OD maps by J. Wood et al.

The starting point of this visualization is a classic OD matrix.

migration_raw_data

For my previous flow maps, I already converted this data into a more GIS-friendly format: a Geopackage with lines and information about the origin, destination and strength of the flow:

migration_attribute_table

In addition, I grabbed state polygons from Natural Earth Data.

At this point, we have 72 flow features and 9 state polygon features. An ordinary join in the layer properties won’t do the trick. We’d still be stuck with only 9 polygons.

Virtual layers to the rescue!

The QGIS virtual layers feature (Layer menu | Add Layer | Add/Edit Virtual Layer) provides database capabilities without us having to actually set up a database … *win!*

Using a classic SQL query, we can join state polygons and migration flows into a new virtual layer:

virtual_layer

The resulting virtual layer contains 72 polygon features. There are 8 copies of each state.

Now that the data is ready, we can start designing the visualization in the Print Composer.

This is probably the most manual step in this whole process: We need 9 map items, one for each mini map in the small multiples visualization. Create one and configure it to your liking, then copy and paste to create 8 more copies.

I’ve decided to arrange the map items in a way that resembles the actual geographic location of the state that is represented by the respective map, from the state of Vorarlberg (a proud QGIS sponsor by the way) in the south-west to Lower Austria in the north-east.

To configure which map item will represent the flows from which origin state, we set the map item ID to the corresponding state ID. As you can see, the map items are numbered from 1 to 9:

small_multiples_print_composer_init

Once all map items are set up, we can use the map item IDs to filter the features in each map. This can be implemented using a rule based renderer:

small_multiples_style_rules

The first rule will ensure that the each map only shows flows originating from a specific state and the second rule will select the state itself.

We configure the symbol of the first rule to visualize the flow strength. The color represents the number number of people moving to the respective district. I’ve decided to use a smooth gradient instead of predefined classes for the polygon fill colors. The following expression maps the feature’s weight value to a shade on the Viridis color ramp:

ramp_color( 'Viridis',
  scale_linear("weight",0,2000,0,1)
)

You can use any color ramp you like. If you want to use the Viridis color ramp, save the following code into an .xml file and import it using the Style Manager. (This color ramp has been provided by Richard Styron on rocksandwater.net.)

<!DOCTYPE qgis_style>
<qgis_style version="0">
  <symbols/>
    <colorramp type="gradient" name="Viridis">
      <prop k="color1" v="68,1,84,255"/>
      <prop k="color2" v="253,231,36,255"/>
      <prop k="stops" v="0.04;71,15,98,255:0.08;72,29,111,255:0.12;71,42,121,255:0.16;69,54,129,255:0.20;65,66,134,255:0.23;60,77,138,255:0.27;55,88,140,255:0.31;50,98,141,255:0.35;46,108,142,255:0.39;42,118,142,255:0.43;38,127,142,255:0.47;35,137,141,255:0.51;31,146,140,255:0.55;30,155,137,255:0.59;32,165,133,255:0.62;40,174,127,255:0.66;53,183,120,255:0.70;69,191,111,255:0.74;89,199,100,255:0.78;112,206,86,255:0.82;136,213,71,255:0.86;162,218,55,255:0.90;189,222,38,255:0.94;215,226,25,255:0.98;241,229,28,255"/>
    </colorramp>
  </colorramps>
</qgis_style>

If we go back to the Print Composer and update the map item previews, we see it all come together:

small_multiples_print_composer

Finally, we set title, legend, explanatory texts, and background color:

migration

I think it is amazing that we are able to design a visualization like this without having to create any intermediate files or having to write custom code. Whenever a value is edited in the original migration dataset, the change is immediately reflected in the small multiples.

Last time, I wrote about the little details that make a good flow map. The data in that post was made up and simpler than your typical flow map. That’s why I wanted to redo it with real-world data. In this post, I’m using domestic migration data of Austria.

Raw migration data

Raw migration data, line width scaled to flow strength

With 9 states, that makes 72 potential flow arrows. Since that’s too much to map, I’ve decided in a first step to only show flows with more than 1,000 people.

Following the recommendations mentioned in the previous post, I first designed a basic flow map where each flow direction is rendered as a black arrow:

migration_basic

Basic flow map

Even with this very limited number of flows, the map gets pretty crowded, particularly around the north-eastern node, the Austrian capital Vienna.

To reduce the number of incoming and outgoing lines at each node, I therefore decided to change to colored one-sided arrows that share a common geometry:

migration_twocolor

Colored one-sided arrows

The arrow color is determined automatically based on the arrow direction using the following expression:

CASE WHEN
 "weight" < 1000 THEN color_rgba( 0,0,0,0)
WHEN
 x(start_point( $geometry)) - x(end_point($geometry)) < 0
THEN
 '#1f78b4'
ELSE
 '#ff7f00'
END

The same approach is used to control the side of the one-sided arrow head. The arrow symbol layer has two “arrow type” options for rendering the arrow head: on the inside of the curve or on the outside. This means that, if we wouldn’t use a data-defined approach, the arrow head would be on the same side – independent of the line geometry direction.

CASE WHEN
 x(start_point( $geometry)) - x(end_point($geometry)) < 0
THEN
 1
ELSE
 2
END

Obviously, this ignores the corner case of start and end points at the same x coordinate but, if necessary, this case can be added easily.

Of course the results are far from perfect and this approach still requires manual tweaking of the arrow geometries. Nonetheless, I think it’s very interesting to see how far we can push the limits of data-driven styling for flow maps.

Give it a try! You’ll find the symbol and accompanying sample data on the QGIS resource sharing plugin platform:

resourcesharing_flowmap

In my previous post, I shared a flow map style that was inspired by a hand drawn map. Today’s post is inspired by a recent academic paper recommended to me by Radoslaw Panczak  and Thomas Gratier :

Jenny, B., Stephen, D. M., Muehlenhaus, I., Marston, B. E., Sharma, R., Zhang, E., & Jenny, H. (2016). Design principles for origin-destination flow maps. Cartography and Geographic Information Science, 1-15.

Jenny et al. (2016)  performed a study on how to best design flow maps. The resulting design principles are:

  • number of flow overlaps should be minimized;
  • sharp bends and excessively asymmetric flows should be avoided;
  • acute intersection angles should be avoided;
  • flows must not pass under unconnected nodes;
  • flows should be radially arranged around nodes;
  • quantity is best represented by scaled flow width;
  • flow direction is best indicated with arrowheads;
  • arrowheads should be scaled with flow width, but arrowheads for thin flows should be enlarged; and
  • overlaps between arrowheads and flows should be avoided.

Many of these points concern the arrangement of flow lines but I want to talk about those design principles that can be implemented in a QGIS line style. I’ve summarized the three core ideas:

  1. use arrow heads and scale arrow width according to flow,
  2. enlarge arrow heads for thin flows, and
  3. use nodes to arrange flows and avoid overlaps of arrow heads and flows

This slideshow requires JavaScript.

To get started, we can use a standard QGIS arrow symbol layer. To represent the flow value (“weight”) according to the first design principle, all arrow parameters are data-defined:

scale_linear("weight",0,10,0.1,3)

To enlarge the arrow heads for thin flow lines, as required by the second design principle, we can add a fixed value to the data-defined head length and thickness:

scale_linear("weight",0,10,0.1,1.5)+1.5

arrow_head_thickness

The main issue with this flow map is that it gets messy as soon as multiple arrows end at the same location. The arrow heads are plotted on top of each other and at some point it is almost impossible to see which arrow starts where. This is where the third design principle comes into play!

To fix the overlap issue, we can add big round nodes at the flow start and end points. These node buffers are both used to render circles on the map, as well as to shorten the arrows by cutting off a short section at the beginning and end of the lines:

difference(
  difference(
    $geometry,
    buffer( start_point($geometry), 10000 )
  ),
  buffer( end_point( $geometry), 10000 )
)

Note that the buffer values in this expression only produce appropriate results for line datasets which use a CRS in meters and will have to be adjusted for other units.

arrow_nodes

It’s great to have some tried and evaluated design guidelines for our flow maps. As always: Know your cartography rules before you start breaking them!

PS: To draw a curved arrow, the line needs to have one intermediate point between start and end – so three points in total. Depending on the intermediate point’s position, the line is more or less curved.

The QGIS map style I want to share with you today was inspired by a hand-drawn map by Philippe Rekacewicz that I saw on Twitter:

The look reminds me of conveyor belts, thus the name choice.

You can download the symbol and a small sample dataset by adding my repo to the QGIS Resource Sharing plugin.

resourcesharing_conveyor

The conveyor belt is a line symbol that makes extensive use of Geometry generators. One generator for the circle at the flow line start and end point, respectively, another generator for the belt, and a final one for the small arrows around the colored circles. The color and size of the circle are data defined:

conveyor_details

The collection also contains a sample Geopackage dataset which you can use to test the symbol immediately. It is worth noting that the circle size has to be specified in layer CRS units.

It’s great fun playing with the power of Geometry generator symbol layers and QGIS geometry expressions. For example, this is the expression for the final geometry that is used to draw the small arrows around colored circles:

line_merge( 
  intersection(
    exterior_ring( 
      convex_hull( 
        union( 
          buffer( start_point($geometry), "start_size" ),
          buffer( end_point($geometry), 500000 )
        )
      )
    ),
    exterior_ring( 
      buffer( start_point( $geometry), "start_size" )
    )
  )
)

The expression constructs buffer circles, the belt geometry (convex_hull around buffers), and finally extracts the intersecting part from the start circle and the belt geometry.

Hope you enjoy it!

It’s holiday season, why not share one of your own symbols with the QGIS community?

In the fist two parts of the Movement Data in GIS series, I discussed modeling trajectories as LinestringM features in PostGIS to overcome some common issues of movement data in GIS and presented a way to efficiently render speed changes along a trajectory in QGIS without having to split the trajectory into shorter segments.

While visualizing individual trajectories is important, the real challenge is trying to visualize massive trajectory datasets in a way that enables further analysis. The out-of-the-box functionality of GIS is painfully limited. Except for some transparency and heatmap approaches, there is not much that can be done to help interpret “hairballs” of trajectories. Luckily researchers in visual analytics have already put considerable effort into finding solutions for this visualization challenge. The approach I want to talk about today is by Andrienko, N., & Andrienko, G. (2011). Spatial generalization and aggregation of massive movement data. IEEE Transactions on visualization and computer graphics, 17(2), 205-219. and consists of the following main steps:

  1. Extracting characteristic points from the trajectories
  2. Grouping the extracted points by spatial proximity
  3. Computing group centroids and corresponding Voronoi cells
  4. Deviding trajectories into segments according to the Voronoi cells
  5. Counting transitions from one cell to another

The authors do a great job at describing the concepts and algorithms, which made it relatively straightforward to implement them in QGIS Processing. So far, I’ve implemented the basic logic but the paper contains further suggestions for improvements. This was also my first pyQGIS project that makes use of the measurement value support in the new geometry engine. The time information stored in the m-values is used to detect stop points, which – together with start, end, and turning points – make up the characteristic points of a trajectory.

The following animation illustrates the current state of the implementation: First the “hairball” of trajectories is rendered. Then we extract the characteristic points and group them by proximity. The big black dots are the resulting group centroids. From there, I skipped the Voronoi cells and directly counted transitions from “nearest to centroid A” to “nearest to centroid B”.

(data credits: GeoLife project)

From thousands of individual trajectories to a generalized representation of overall movement patterns (data credits: GeoLife project, map tiles: Stamen, map data: OSM)

The resulting visualization makes it possible to analyze flow strength as well as directionality. I have deliberately excluded all connections with a count below 10 transitions to reduce visual clutter. The cell size / distance between point groups – and therefore the level-of-detail – is one of the input parameters. In my example, I used a target cell size of approximately 2km. This setting results in connections which follow the major roads outside the city center very well. In the city center, where the road grid is tighter, trajectories on different roads mix and the connections are less clear.

Since trajectories in this dataset are not limited to car trips, it is expected to find additional movement that is not restricted to the road network. This is particularly noticeable in the dense area in the west where many slow trajectories – most likely from walking trips – are located. The paper also covers how to ensure that connections are limited to neighboring cells by densifying the trajectories before computing step 4.

trajectory_generalization

Running the scripts for over 18,000 trajectories requires patience. It would be worth evaluating if the first three steps can be run with only a subsample of the data without impacting the results in a negative way.

One thing I’m not satisfied with yet is the way to specify the target cell size. While it’s possible to measure ellipsoidal distances in meters using QgsDistanceArea (irrespective of the trajectory layer’s CRS), the initial regular grid used in step 2 in order to group the extracted points has to be specified in the trajectory layer’s CRS units – quite likely degrees. Instead, it may be best to transform everything into an equidistant projection before running any calculations.

It’s good to see that PyQGIS enables us to use the information encoded in PostGIS LinestringM features to perform spatio-temporal analysis. However, working with m or z values involves a lot of v2 geometry classes which work slightly differently than their v1 counterparts. It certainly takes some getting used to. This situation might get cleaned up as part of the QGIS 3 API refactoring effort. If you can, please support work on QGIS 3. Now is the time to shape the PyQGIS API for the following years!

The possibility to easily share plugins with other users and discover plugins written by other community members has been a powerful feature of QGIS for many years.

The QGIS Resources Sharing plugin is meant to enable the same sharing for map design resources. It allows you to share collections of resources, including but not limited to SVGs, symbols, styles, color ramps, and processing scripts.

Using the Resource Sharing plugin is like using the Plugin Manager. Once installed, you are presented with a list of available resource collections for download. You will find that there are already some really nice collections, including nautical symbols, Mapbox Maki Icons, and my Google-like OSM road style.

resourcesharing_maki

By pressing Install, the resource collection is downloaded and you can have a look at the content using the Open folder button. In case of the Mapbox Maki Icon collection, it contains a folder of SVGs:

resourcesharing_folder

Using the new icons is as simple as opening the layer styling settings and selecting the Mapbox Maki Icons collection in the SVG group list:

resourcesharing_styling

Similarly, if you download the OSM Spatialite Googlemaps collection, its road line symbols are added to your existing list of available line symbols:

resourcesharing_roads2

By pressing the Open Library button, you get to the Style Manager where you can browse through all installed symbols and delete, rename, or categorize them.

resourcesharing_roads

The Resource Sharing plugin was developed by Akbar Gumbira during this year’s Google Summer of Code. The full documentation, including instructions for how to share your own symbols with the community, is available at www.akbargumbira.com/qgis_resources_sharing.

%d bloggers like this: