Sankey Diagrams

Just before the holidays a new version of e!Sankey became available. I had some time to play around with the new 2.5 release. Not that many new functions, it seems, but apparently some bug fixes and smaller improvements. The main new feature is the interface langauges in French, Portuguese and Spanish. The entry in my Sankey diagram software list has been updated accordingly.

End-of-year cleanup of my bookmarks and saved diagrams. There are more goodies, and enough in stock for a couple of posts next year…

Below are three image sections of a Sankey diagram showing Swedish energy flows in 2007. These were published on p. 54/55 in the report ‘Energy in Sweden 2008′ by Swedish Energy Agency (Energimyndigheten). Numbers are in TWh, source for the diagram was Statistics Sweden data.

The first diagram (which is the left part of the diagram) shows the total 624 TWh energy supplied and the energy sources. The second one is the breakdown of energy produced by sectors, and also shows the losses in energy conversion. The third on is a different breakdown of the 404 TWh by energy carrier.

Nothing more to add. They are simple, clean, beautiful.

The Mountain State saw the annual Governor’s Energy Summit in Roanoke on Dec 9, 2009. West Virginia relies heavily on coal, and is a net energy exporter to other U.S. states.

The energy flow Sankey diagram below (created by Marshall University for WV Energy Division) was presented at last year’s energy summit, and is available along with the other presentations here.

Figures are for 2006 in trillion BTU. Energy carriers used in WV are displayed as flows entering from the left. Domestic energy sources are in orange, while imported supplies are in sand color. Overall energy in 2006 was 4,384 trillion BTU. The state exported 81% of the energy (blue) and consumed 19% within (836 trillion BTU). A breakdown of doemstic consumption by sectors is shown in pink.

This Sankey diagram looks nice, but violates the basic rule for Sankey diagrams: flows have to be to scale among each other. The magnitude of the stacked orange arrows (representing 4100 trillion BTU West Virginia production) should be 14.5 times larger than that of the sand color flow (representing 283.86 trillion BTU), however it is only about 9 times larger, overemphasizing external supply (or underrepresenting domestic energy supply).

Or, compare the two arrows fro “crude oil” (10.14) and “natural gas” (230.12). The latter should have 23 times the width of the other… The blue arrow for “international raw coal” (392) looks approximately as wide as that of the blue “natural gas” (106). The scale might still be somehow OK for the base of the arrow, but as the arrow becomes thinner towards the head, the 4:1 ratio is definitely not supported any more.

I have posted several Sankey diagrams depicting the energy flows of countries. At least California and West Virginia have published state energy flow Sankey diagrams. I was quite excited to discover the two metropolitan energy flow Sankey diagrams shown below in this publication. They are for Toronto and Helsinki and show energy flows in 1988 in these communities.

The two diagrams show energy consumption and use in Toronto (above) and Helsinki (below). Even though the absolute figures in GWh are given, one shouldn’t directly compare them. A per capita basis would be fairer (Toronto had a population of 2.5 mio in 2006, more than 5 mio. in the metro area, while Helsinki had 580.000 inhabitants in 2008 in the city, 1.3 mio in the greater Helsinki area — Toronto is today 4.5 times larger than Helsinki). Both are “cold-climate municipalities”.

The publication calculates a ‘community energy efficiency’ of 50% for Toronto and 68% for Helsinki. “A comparison of the two municipalities reveals that Helsinki significantly improved its efficiency by using the waste heat that is produced by local coal power plants to warm 90% of the buildings and homes in Helsinki. Further analysis has demonstrated that Helsinki’s energy system was able to achieve its overall level of 68% efficiency because the city’s compact land-use pattern made investments in energy-saving infrastructure, such as district heating and public transit, economically viable.”

Does anybody know other metropolitan energy flow studies? I am aware of research activities in the field of urban material flow accounts or urban metabolism (e.g. Lisbon) but have to check whether they show Sankey diagrams in their publications.

The Visio Guy had another cool Sankey diagram on his blog last week. Credits go to Chris Webb of Woodland Trust, who created this using the line thickness option rather than pre-wired shapes.

The diagram has a left-to-right orientation and shows the different sources of money received by the trust. The types of funds (e.g. grants, legacies, direct marketing) are grouped together by colors. Flows have percentile values, rather than absolute ones. I am not sure what the boxes labeled “Sys” are, but the colors change. All flows merge into the box “Finance” which has a subgroup “Sales Ledger”.

The flow bands between most of the nodes have a nice soft curving. This is why some people do refer to Sankey diagrams as spaghetti diagrams.

If you are using Visio, you can download this diagram and look how it has been done. Nice work! I hope to see more of these Sankey diagrams done in Visio….

This website on energy optimized building features a case study of a students dorm building at Wuppertal, Germany. [On a side note: it is quite striking that when I browse for new Sankey diagrams on the web, I come across them on German websites. Herr Sankey could well have been a German...]

The two Sankey diagrams shown along with the article visualize the heat energy flows before and after the optimization measures taken.

Flows are in kWh per square metre per year. The first diagram breaks down heat consumption into heat used for room heating and for warm water preparation. In the second a new heating system has been installed providing heat mostly over the air circulation system. Only some static radiators remain. [Thanks Thomas for helping me with the translation/explanation].

Apart from the use of a fancy color, the Sankey diagrams are well done and comprehensible. Flows are to scale.

In 2008, the Cartographia blog started a post series called ‘Monday’s with Minard’. Some people consider Charles Joseph Minard the first to use arrow magnitude in his diagrams to represent quantities. (As a consequence, this means that Sankey diagrams would have to be renamed to Minard diagrams!).

What differentiates Minard maps from Sankey diagrams is that Minard’s fine works always have a geographical relation. The most famous one is his Map of Napoleon’s march to Moscow published in 1869. This “carte figurative des pertes successives en hommes de l’Armée Francaise dans la campagne de Russie 1812-13″ shows number of men (as width of arrows), geographic movement of the troops on the map both for invasion as well as for retreat, as well as time and temperature on a separate scale.

Cartographia blog has some other nice examples, two of which are shown here:

The first shows migration patterns across the globe. Arrows do not have an arrow head but the country of emigration is color coded. The outline of the countries is distorted to accomoadate large flows connected to them. For a detailed description please consider reading the original blog post. This map is similar to the one I showed in this post.

The other is a flow map for wool and cotton for the years 1858 and 1861. “Blue represents cotton and wool from the United States, the orange from British territories in South Asia … One millimeter represents 5,000 tons of cotton or wool.”. As one can see on the 1861 map, cotton imports from Asia have increased dramatically. See the description of the map in the blog post on Cartographia blog.

See all Monday’s with Minard posts here. There has been no activity on the blog since June 2008. I hope to see more of these posts some day.

A group of graduate students form the Center of Industrial Ecology at Yale University in 2005/2006 researched the material flows on the “Big Island of Hawai’i”. Their research report (which can be found on the website of the Kohala Center) shows two Sankey diagrams, one of which is shown here.

Material Flow Accounting (MFA) “is the study of material flows on a national or regional scale. It is therefore sometimes also referred to as regional, national or economy-wide material flow analysis.” (Wikipedia). MFA is a research field in industrial ecology. As the authors of the report write,

Using an island as a unit of analysis is valuable both to the researcher and to those interested in the sustainability of the island itself. The researcher benefits from the island’s clear boundaries (most often defined by a surrounding water body) and a relative advantage in data collection provided by the fact that borders are monitored. Material flows are therefore relatively easier to understand on islands than in larger, more complex non‐island systems.

I have been posting about the use of Sankey diagrams in MFA before, and with few exceptions (Material Flow Sankey Diagram of Japan), have found that examples of Sankey diagrams for national MFA accounts typically are limited to selected bulk materials (e.g. biomass in Switzerland, gold flows in the U.S.).

Flows in the above Sankey diagram for the island of Hawai’i are in gigagrams (kilotonnes) and refer to the year 2005. Inputs are shown on the left side, and the fate of those inputs can be seen as exits to the right. More than 75% of the material flows are imported from off the island, the majority of these flows (57%) are construction materials. Consequently, road and building construction are the largest net addition to stocks with almost 2,000 kilotonnes.

The Sankey diagram has some minor flaws, regarding scale of the flows. Look for example to the division of the landfilled waste arrow (422 kt) into three almost equal portions, which are supposed to represent 79.3 kt, 125.1 kt and 217.7 kt. Also the width of the volcanic rock input flow (429.7 kt) is about four times the width of the machinery input flow (representing 200.2 kt). Still, I think it is a good Sankey diagram, and I wouldn’t mind joining the research group on their next visit to the islands…