Not always is it meaningful to map everything that has a coordinate. Recently visualisations of road fatalities appeared on the web that used public data to show every death on every road in a relatively large period on a map. The amount of data concerned resulted in a map that looks similar to a visualisation of the traffic density, which highly correlates with the population density. Hence, this map does not really provide a new perception. I’ll give following map by the BBC as an example:
A smart but simple visualisation by FlowingData breaks the data down into seasonal variations instead of spatial ones. This way new interesting trends become visible like the higher number of accidents on weekends or through the summer months.
In case one wants to use the coordinates included in the data, it would make sense to combine it with other spatial information such as traffic density on certain road sections. This way spatial centers of gravity for road accidents could become visible.
I actually wanted to post this a week ago but due to lazy Easter holidays it comes a bit delayed. The topic however is still in everybody’s minds. Gregor Aisch has produced an interesting visualisation for the ZEIT magazine. It shows that in a densily populated country like Germany many people would be affected by a nuclear disaster like in Fukushima. Japan is densily populated too, however there the people seem to be more concentrated in large cities than in Germany. I heard about 80.000 people were living in the evacuation zone near Fukushima. In Germany an evacuation radius of 20km would affect 856.000 people for the plant in Neckarwestheim alone and 1.7 million around Philipsburg in case of a 30km radius. 5% of the country’s total population live within 20km of a nuclear power plant, 12% within 30km and 51% within 80km.
The south west of Germany is an earthquake risk area. However no strong quakes like in Japan have been recorded there. The strongest was 6.5 on Richter scale in Basel (CH) in 1356.
I had the topic refugees and Europe before when pointing to a map by Philippe Rekacewicz about the Schengen border. Now it is on the agenda again since the fortress Europe worries about the growing number of “intruders” and the lack of a consensus on how to deal with them (http://www.bbc.co.uk/news/world-europe-13109631). The data journalism specialist OWNI has a nice interactive map on that topic:
Central and Western Europe are freezing these days. Such widespread frost and snow at this time of the year is quite unusual and creates the appearently inevitable chaos on road, rail and in the air. Since even Great Britain is hit hard by this early winter the guardian’s Datablog came up with a list of temperatures measured throughout Britain on December 20th. The temperatures are quite impressing – -18.7°C in England is quite something.
The plot of the measuring stations on google maps on the article’s side to me is nothing but playing around. What we want here is a decent map that visualizes the distribution of the temperature. One comment below the article linked to an application called bime. The “Heat Map” seems to me like a visualisation of the density of the measuring stations. At least one can use a slider on the right side to limit the range of temperatures displayed, but a comprehensive overview is not given. When switching the rendering mode to graduated circles the visualisation does not make sense at all.
Another google mashup that enables querying of the data is also linked: http://homepage.ntlworld.com/keir.clarke/web/cold.htm. I am not really satisfied with that either. The shown visualisations don’t represent the theme temperature as the continous phenomenon that it is.
The authors encourage their readers to download the data as spreadsheet and let them know what they did with it. So here it comes:
My suggested solution is the following:
- download the data as spreadsheet
- import it in Quantum GIS by using “Add Delimited Text Layer”
- add a shapefile with outlines of Britain; for some reason my outlines do not exactly match with the plotted stations, since some of them are clearly placed in the ocean although they probably represent coastal towns. Since this is a rather quick hack I did not fix that.
- use the plugin function “contours” to interpolate the temperature values; take the Layer with the stations as input, select “min temp air temp overnight” as Data Field; select filled contours; I used 8 classes to classify the range of temperatures occuring
- now one should give meaningful colours to the contours; in the newly created layers with the temperature polygons select properties and go to symbology; select graduated symbol. The result should be a colour range of cold colors of different shades of blue, however at the freezing point mark there should be a transition to another colour, prefereably green to visually differentiate between freezing and non-freezing areas.
- I decided to label some of the measuring stations with their corresponding temperature. In the properties of the layer I selected label and ticked “Label only selected features”. In the main view I then selected some stations with some distinct values and made sure there is no large area without a label.
Here is the result – a rather quick hack. It is not a nice finished map yet but shows in which direction it should go. I feel that the values sometimes were not matched to the classes perfectly. The plus 1.2° in Cornwall for instance should be in the green class according to my classification. So there is potential to play with it further.
A beautiful interactive infographic by Gregor Aisch visualizing the donations to political parties in Germany can be found on the blog vis4.net.
This is especially interesting since the influence of lobbyism on German politics has become more and more clear recently. More transparency in this issue is crucial. Germany is one of the few countries in the world that has not ratified the United Nations Convention on Corruption (UNCAC) which is mainly because of a lack of law enforcement in the field of bribery of members of parliament.