The below Sankey diagrams both show wood biomass flows for Finland for the year 2013.

The first one was published in the report VTT Technology 237 ‘Sustainability of forest energy in Northern Europe’ by researchers from VTT Technical Research Centre of Finland and Natural Resources Institute Finland (Luke).


Authors of this figure are Eija Alakangas and Janne Keränen. The diagram is oriented top-to-bottom and shows how the 104.4 Mm³ of round wood that grew in Finnish woods in 2013 were used. Basically there are two (three) main pathways, with a lot of arrows branching out to depict certain uses. 38.3 Mm³ of round wood was used in pulp industry, 26.2 Mm³ in the mechanical wood industry. Another 9.5 Mm³ of wood is used directly for energy generation.

The second Sankey diagram seems to be a remake of the above. It was published in a VTT Research Report on ‘Cascading use of wood in Finland – with comparison to selected EU countries’ by Laura Sokka, Kati Koponen, Janne T. Keränen.


Here the overall orientation is left-left-to right. The color scheme seems similar. There are some minor differences in the energy use part (orange and dark red arrows).

The first diagram has some images and comes across a little more playful than the second one. Although they depict the same data, I perceive them quite differently.
Is it due to the scaling or the vertical vs. horizontal orientation? Let me know your impression in the comments please.

This week the global plastics flows topic made the news and social media with the publication of the EU Plastics Strategy and Chancellor Philip Hammond presenting the United Kingdom’s plan for tackling plastic waste.

Ellen MacArthur Foundation has long been active in research and awareness building in this field. It aims at supporting a transition to a circular economy. The foundation tweeting under @circulareconomy contributed this Sankey diagram. It is from a 2016 report they produced together with the World Economic Forum and McKinsey.

The Sankey diagram shows indeed, that “today, plastic packaging material flows are largely linear”. This beautifully crafted diagram had already caught my attention back in 2016 when I first saw it.

However, I had this subtle feeling that something was wrong here. Not regarding the content or the data … but rather that something wasn’t OK in the Sankey diagram, Just my gut feeling. Now, seeing the Sankey diagram again in the above tweet this week, I finally sat to quickly do a remake of this Sankey diagram. Here it is:

I stuck to the original layout and design as closely as possible, using the same color codes and even the white all caps font. While transfering the numbers (all percentage values, so no issue there), it immediately became clear to me what caused my irritation. Can you identify it yourself by comparing the two pics?

Won’t give it away now and wait for your comments. Will post the answers to this small ‘spot-the-difference contest’ here next week.

[Edit 24 Jan] Blog reader ‘First!’ was the first to comment and point out that the 2% recycling flow does not seem to be to scale (i.e too wide / overemphasized) in the Sankey diagram published by Ellen MacArthur Foundation, and possibly the same issue with the two arrows representing 14% each.

‘Dubbel’s Handbook of Mechanical Engineering’ could be considered a bible for mechanical engineering students in Germany. Despite its 900 pages it is still called a pocket book (‘Dubbel – Taschenbuch für den Maschinenbau’) in German quite euphemistically. Since it was first published in 1914 by Heinrich Dubbel it has seen some 24 editions and roughly 920,000 copies sold. Since 1994 it is also available in an English translation from Springer Publishers.


Someone challenged me, if I could do the above figure from Dubbel’s book (“Wärmestrom in einer Kesselanlage”, heat flow in a boiler system). I did various copies and here are the two I like best:

In this version losses are shown in grey with a gradient to dark grey.

The other one sticks closer to the original with the hatch pattern on the arrows representing losses. I had to fill the nodes since the contrast between the colored arrows and the hatched arrows was just too harsh.

I confess I couldn’t do the labels with prime marks and subscript directly in e!Sankey. So I did them in Word, created tiny images and rotated them as work around. Later I found out that the whole image was probably originally intended to be displayed vertically, but rotated to the left only to save space in the book.

Anyway … a fun challenge. I hope you like the result. Let me know your opinion.

The International Energy Agency (IEA) is a good source for reports on energy, both with a focus on global energy, but also breaking it down to the national level. I have featured their Sankey diagram website that allows to access national energy balances for many countries in this post back in 2013.

Browsing through their reports also sometimes reveals Sankey diagram gems. In their report on ‘Tracking Industrial Energy Efficiency and CO2 Emissions’, however, I found the diagrams on aluminium, steel, pulp/paper and petroleum not particularly sexy.

This is a schematic block diagram. Arrows are labeled with the quantity in Mt/year.

I decided to redo this as a Sankey diagram, maintaining the general structure of the original diagram. The width of the Sankey arrows immediately exhibit where most of the mass (crude oil) is…

I chose three colors: blue for the actual products from petrochemical industry, yellow for recycling streams and losses, purple for the precurors or feedstock (I actually thought I should do away with these, since the ‘hydrogen energy’ flow gave me some headache…). Also decided that the head of the arrow representing 115 Mt/year of post-consumer waste leading towards (!) net additions to stock in the original diagram is erroneous and thus turned the arrow around.

Didn’t spend much time on graphic aspects or fine tuning. I am sure this can be done quite nicely. But even like this I think a Sankey diagram is the better way to get the message across.

Browsing through the blogs on data visualization and infographics (check my blogroll) I often find inspiration in Nels’ MFA diagrams. From time to time I like to beef up the skinny MFA diagram skeletons a bit by converting them into Sankey diagrams. At the same time, by translating the numbers into Sankey arrows one gets a better idea what the main (mass) flows are.

This is a MFA diagram on Iron and Steel Flows in the European Union in 2000 as found in this post. Original data is from a 2008 OECD study, flows in Mt.

The description of the diagram says: “A study of iron and steel flows in 2000 in the European Union showed that an input of about 120 Mt of iron ore (of which 98 Mt was imported) yielded 98 Mt of primary crude steel (i.e. produced directly from iron ore and coke). A further 65 Mt, representing 40% of total crude steel production, were produced as secondary crude steel, produced from scrap steel.”

I did a first quick version of the flows as Sankey diagram, trying to stick very much to the layout of the original diagram. All nodes are the same size and more or less located at the position of the master. It already shows that the main steel flows: iron ore imported into the European Union, and steel scrap being recycled within the EU. Export of semi-finished steel products from the EU to the Rest of World (52 Mt) almost balanced with 47 Mt of semi-finished steel products imported into the EU.

I tried to improve the diagram by removing the three nodes ‘New Scrap’, ‘Prompt Scrap’ and ‘End of Life products’ since there is no transformation of these flows at the nodes (also no change in quantity). Further I reduced the size of some boxes and dragged the ‘Semi-finished Products’ (Rest of World) box closer to the ‘Finished Steel Products’ (European Union) box to avoid crossing streams. Wherever possible I try to avoid diagonal arrows.

The final result also has the Rest of World and European Union grouping. I am not to happy with the colors though.

Your thoughts?

Phosporus in the natural environment and the food chain has been a topic of several posts on my blog. So it didn’t come as a surprise to find yet another diagram on phoshphorus flows over at Nels’s MFA Diagram blog (one of the blogs I follow closely, see blogroll).

MFA diagrams have their focus on the nodes and the build-up of stocks. Sometimes they get a touch of Sankey diagram with the arrows having different magnitudes. The MFA diagram below is for phosphorous flows in China 2008 (original source: Min Qiao, Yuan-Ming Zheng, Yong-Guan Zhu, 2011. Material flow analysis of phosphorus through food consumption in two megacities in northern China). Values are in tonnes.


(click image to enlarge)

We can detect arrows with three different brush widths (my guess is 1px, 2px and 4 px), each standing for a value range into which the actual flow quantity falls. This may, however, bes somewhat misleading when having a quick glance at the diagram.

I quickly “translated” the above diagram to a Sankey diagram with flow values being actually to scale.


(click image to enlarge)

Here it is quite clear where the major phosphorus flows are located (from food production via urban consumption to sewage treatment plant and solid waste disposal: 2923 out of 5374 tons end up here). The other flows are comparatively small, with the phoshporous flow going directly to the aquatic system worth a mention. Two small flows in the center of the diagram are negligible, they are in fact so tiny in comparison to the major flows that they even don’t show up (or just as a hairline) here.

I have therefore added a minimum width of 1 px for small flows so that the annual 17 tons from urban consumption and the 1.9 tons from rural consumption to the solid waste disposal are at least visible (albeit not to scale with the other flows any more).


(click image to enlarge)

Final phosphorous sinks are solid waste disposal (landfill?) and the aquatic system.

Below is an example of a Sankey diagram showing a nitrogen metabolism. The original diagram is from a Japanese publication (‘White Paper on Quality of the Environment in Japan 1994’), even though the diagram represents nitrogen loads (in 1000 tons N) in the Netherlands in 1990.

A pimped version of this diagram can be found in the e!Sankey download gallery. I don’t like the color very much, but the overall aspect of the diagram is much better than in the b/w version, I think.


They seemed to have struggled with inconsistencies in the original diagram, as an annotation suggests. Also the denitrification due to accumulation in the bottom sediment, or nitrogen ending up in durable goods (shown in black in the original) are not represented in the remake. For the rest it pretty much sticks to the original.

Thinking about a new tag “pimp my Sankey”…