Hi Chris,

I haven't done this and am not aware of anyone who has. However, scanning the list of web services I can see it is theoreticaly possible using the One Species Data service. Taking heed of the warningsone might retrieve records for a species, a year at a time, and aggregate counts grouped by date to produce data to draw a phenology plot. You could also draw multiple plots to show changes in phenology with time.

Since that would be quite time consuming you would probably want to create a cache of results and only refresh this periodically. It might even be the case that you build the cache as an entirely separate process from your website. It is a job that could be much more efficiently done on the NBN servers than yours it seems to me. Why haven't they done this?

It prompts me to wonder about the validity of such a plot. Certainly, in the same way that you have to interpret distribution maps carefully since they can be a map of where recorders live, a phenology plot could be a plot of when recorders go out. The better recorded a species is, the less of an issue this would become, as always.

It would be interesting to know if there others who would like to derive phenology plots from NBN data.

Jim Bacon.

Thanks for the replies. I agree, it seems something that the NBN systems should be capable of, as they have sufficient  data to produce basic plots, as they do with maps.

I don't think that the limitations of phenology graphs should influence NBN's ability to return them ... even mapping has its limitations too, such as a lack of precision in the grid references. If the client website could obtain the number of records in the system then the client website itself could decide whether to ask for a phenology graph or not - and the NBN query could remove duplicate records fairly easily.

The NBN has a lot of data - we just need the tools to extract the common graphs (phenology/abundance) and we can display them and make use of them. Phenology is becoming a very common thing that people want to analyse, what with climate change. :)  As you said Jim, extracting the data to a client's database and then doing the work there seems to be a very heavy-handed way to do it.

I too was banned! Probably the automatic software spotted I was rambling inconsequentially. I won't let a little thing like that stop me however...

Don't forget to consider how to display the information once you get it - displaying seasonality is a hobby horse of mine, for species with long/all year long seasons I personally think it is better to use radar plots than bar/line graphs like I did here.

Another thing that bothers me and in all likelihood no one else in the world - I dislike people calling them "phenology graphs"! I have always understood the term phenology to be the study of the *change" in seasonal patterns - so it's only a phenology graph if the plot is comparing seasonal data from more than one time period (which the "phenology" in Recorder 6 doesn't). Dictionaries seem to agree with me, however I see that Wikipedia thinks the use of the term as meaning seasonal occurrence is OK, so perhaps it's just one of those evolving terms I should stop being pedantic about. :rolleyes:

This doesn't of course help you get the data you need.  I do agree aggregate data like this would be a useful thing to be able to access via NBN webservices - especially if you could restrict it by criteria such as datasets as Charlie suggests. :cool: There would doubtless be other uses.

Also nothing wrong with plots that show recording bias, as long as you are aware of it and question the results. I looked at the phenology of bumblebee records using the data at Kent records centre - it seemed to show a recent increase in winter activity, but was that just because an upsurge in recording meant more people were sending winter records in? IMO ad-hoc records are for producing questions, not answers.