Digging for Data

11 August 2014

Region: Headquarters

Have you ever wanted to be recognised as a trend-setter? Do you consider yourself to be a dedicated follower of fashion as illustrated by the photographs you post on the web? Maybe you don't, but other people might think so.

You may have heard the term 'data mining'. This is basically the process of reading large amounts of data (which includes text and images), analysing them and deriving more data as a result. The process is so significant that a new addition to UK copyright law has arrived which permits it under certain circumstances, since the analysis will involve copying into a computer memory.

The fashion example is quite interesting, as by analysing what a large number of people like in terms of their clothing style can be a gateway to predicting changes in what the population as a whole would like to buy. In the past this was done by reading magazines and observing people in the street.  (Some people have a knack of leading the stylistic pack. They are the bellwethers and I heartily recommend you read Connie Willis's novel 'Bellwether' to explore this concept.) What we know of historic fashions often come from old illustrations and photographs. Now the web provides the hunting ground for trends. This has been going on for a number of years: the web site Polyvore explicitly asks its users to experiment  by creating collages of their favorite outfits and furniture, laid out as they might be in a fashion magazine. They've been doing this since 2011. Similarly, what people buy or listen to on music sites gives a good basis for predicting trends in music.

Andy Finney is an RPS member and also writes for http://www.infrared100.org/. - See more at: http://www.rps.org/blogs/2014/may/copyright-on-thumbnails#sthash.GCzRWuYu.dpuf

More recently, some companies have been trawling the internet to find photographs that include fashion items. Their software can analyse an image and identify clothing in it, determining colour and styles, and this information also informs a commercial understanding of future trends. If you have posted photographs of people on the web, such as through social media, then your photographs can be part of this vast database of images.

I'm not suggesting that there is anything unethical about this (although the data mining exception is granted only for the purpose of non-commercial research); and I don't think you should be worried about it. I mention it simply because it is a use of your images of which you may not be aware. I believe that when we post images on the web we do so with an implicit understanding of what might happen to those images. It's known as an implied licence and it's possible we have now to add data mining to those things we expect to happen when we post our photographs.

Andy Finney is an RPS member and also writes for http://www.infrared100.org/.