How do features get rendered?

Hi. I’m new to openstreetmap and flogging through, trying to understand how it works. I’m interested in adding ecological data on the form of JNCC phase 1 data. There is already a proposal to do this.

What I’m wondering is, how does the openstreetmap know how to render stuff? Does the website only render pre-defined feature types? What I want to do is generate a phase 1 layer. My question is, does osm hold information on how the features are rendered, and is there a web page or application that can render my data, or do I need to write software to pull the phase 1 information from osm and render it (either on a web or desktop app) in accordance with the JNCC colour/pattern definitions?

Many thanks for any advice.

David Wallis

Features on the map are pre-rendered, but there are no rendering instructions in OSM (it would defeat the point).

To render something like Phase 1 data you would need to create your own rendering rules: I actually have a set of phase 1 (woodland only) style sheets for QGIS, but for more generic work I would probably choose to use TileMill. QGIS is better as a place to start, but Maperitive is the easiest to use if you want to render directly from an OSM file. (Most approaches involve loading a transformed set of OSM data into a Postgres database).

As an aside, I would avoid, or at least refine the JNCC format as it is very dated and reflects technology constraints of about 25-30 years ago: the letters used as symbols can these days be replaced by much better symbology. Phase 1 is relatively small and requires only a small number of rendering rules.

The existing ideas about tagging habitats (strictly biomes) (e.g., tags for habitat and plant_community) work, but aren’t truly integrated into the overall OSM tagging scheme. The more people who wish to add things like Phase 1 data, the more likely we are to refine it.

There are things which may cause problems: recently felled woodland comes to mind (although in true OSM sense recently_felled=yes will do the job).

For non-British non-ecologists (the bulk of readers of this forum): Phase 1 is a high level ecological classification, which as its name implies should be used as a first step to producing more refined habitat mapping.

For UK specific discussions you might be better of using the talk-gb mailing list particularly as it is much more heavily used than this forum by British mappers.

Thanks for your reply. I’ve installed TileMill and I’m going through the tutorial. I’ve also been using qgis recently and I’m getting up to speed on that.

I’m quite keen to get this up and running and promote it, and hopefully get other people involved. A national phase 1 map (although a massive undertaking) would be a tremendous resource for landscape ecologists. I think that getting a website up and running that can render the features is the first step, so I’ll crack on with that in my spare time (ha ha). I would then aim to promote it in colleges and universities, and try to encourage course leaders to get their students to put their phase 1 maps from coursework on to openstreetmap.

Do you have any links to projects already putting ecological data on to openstreetmap in the UK?

Many thanks for your help,


Unfortunately I dont.

I have the former CCW Phase 1 mapping of Wales, but this has onerous restrictions on the license: it should be under OGL with perhaps OSGB but no more. The problem with these official datasets is that they were a) collected over a long time and b) are not usually maintained. Talking to a friend who was a former employee of CCW highlighted how crowd-sourcing could start with existing data sets and enhance & maintain them.

Another example Nottinghamshire was mapped at Phase 1 in the late 1970s. The data is on paper and has not been maintained. Nottingham was surveyed in the early 1980s by people paid by Manpower Services & has not been maintained.

So to move forward we could do with the following:

  • Move data sets to some kind of Open Licence (Govt., QUANGOs, NGOs like the National Trust & Wildlife Trusts). It would be nice if ecological survey data collected as part of planning applications could be open sourced in some way.
  • Identify the relationship between categories of specific datasets (Phase 1, NVC, EURINE) as has already been done for CORINE and Urban Atlas.
  • Extend and develop tagging as appropriate.
  • Define symbology for QGIS & CartoCss (possibly also Maperitive).

For basic rendering, the Painter’s Algorithm is fine. For providing extracts of data (e.g., in Shapefiles) a much more complex process will be needed.

I suggest placing any CartoCss rules you develop somewhere where they can be shared (OSM Wiki, github etc).