Tuesday, 13 December 2016

Trends in photographic recorder activity

Following up on yesterday's post, I thought it likely that there would be some correlation between some of the trends in species recording and trends in recorder activity. I therefore looked at the numbers of recorders in each year since 2013 and the composition of the dataset in terms of the numbers of records submitted by recorders.

What emerges is a fantastic story of evolving active recorders. The UK Hoverflies Facebook page started in 2013 and at that time was pretty much like other FB pages - lots of posts of photographs, but rarely great runs of photos of different species. By 2016, there were about 30 people who took such an interest that their posts often depicted 10 or more species for a single observation session. Obviously, this translated into vastly more records and individuals who are submitting sizeable datasets. This shift is most heartening because it is this sort of recording that is most likely to generate the sort of data that is needed to detect long-term trends.

The graphs strongly suggest that change in recorder activity is indeed the main factor in some of the relative abundances seen in the species data. I suspect that the transition has now taken place and that from around 2015 onwards the data will be robust enough to detect trends. In that time, there have been a number of other changes. For example, several recorders now submit data as spreadsheets comprising anything from several hundreds to several thousands of records. I have not included these in the analysis because they would complicate the picture. Nevertheless, for long-term trend analysis these sorts of datasets are essential.

If, over time, we see a shift from existing active members submitting records as posts on FB to submitting records as spreadsheets, the dataset for long-term analysis will continue to be strong. Meanwhile, I would hope that there would be newcomers that start to post long runs of photos. Thus we might hope to see a gradual stabilisation of the numbers of records directly extracted from the FB page, and a growth in records submitted as spreadsheets.

For me, the big challenge comes in developing mechanisms for providing regular feedback via this page and via the FB group. That should be possible, but will require a bit of data manipulation. I will have to look at the best way of achieving this.

The data on numbers of records submitted are presented here as two graphs. The first is a simple log2 expression of the numbers submitted. It tells quite a similar story to the second, which is a log2 scale of the percentage of records submitted each year. This latter expression is helpful because the numbers of recorders has dropped from a peak of 1271 in 2014 to 938 in 2016. Much of this decline almost certainly arises because I am not as assiduous in scanning Flickr for new recorders; most of whom would probably only contribute a few records. This, too, is part of the evolution of photographic recording: there is less time to concentrate on the process of searching for recorders and more time needs to be committed to providing feedback!
Figure 1. Numbers of recorders of a scaled range of records submitted (at log 2 scale).

Figure 2. The numbers of recorders, represented as a % of total recorders, against a scaled range of records submitted (at log 2 scale).

No comments:

Post a Comment