Read to Learn

Read to Learn is an entry to the MOSAIC Developer Competition. It is delivered 'as is' and is liable to breakage or removal at any time. Read to Learn is an application designed to suggest possible courses of study based on lists of ISBNs. The concept is that you could upload a list of ISBNs of books you have enjoyed, and Read to Learn will suggest some courses you might like to study. In this way, Read to Learn attempts to facilitate the transition from informal learning to formal learning. You can try it now, or read a bit more detail below.

Select a file which contains a list of ISBNs (of books you have enjoyed), each one on a new line:
Alternatively, to use a supplied list of ISBNs, check this box:

If you check the box to use the supplied list of ISBNs, you are using a list that has been chosen specifically to showoff the functionality of this service. If you'd prefer a more realistic test, but don't have a file of ISBNs to hand, then you can download this file of ISBNs generated from a random RefShare Folder, and then upload it to Read to Learn using this form.

How it works

Read to Learn uses the API very kindly provided by David Pattern to the MOSAIC data set (documented on David's blog). For each ISBN in the uploaded file, Read to Learn checks which courses it is related to and extracts the course 'code'. Currently, all the codes seem to be UCAS codes, and Read to Learn treats them like this. Each code is looked up against the UCAS course catalogue, and a list of relevant courses at institutions across the UK is retrieved and displayed.

If you choose the 'Look for all matches option', the Read to Learn finds displays courses even if you only have a single book in your collection that matches the course. This is likely to give results that are too general to be useful to an individual, especially for works of popular fiction. If you choose the 'Give me close matches only' option, you will only get a recommendation if your collections contains at least 1% of the total items related to that course. Although this latter approach is much more likely to be useful (and could be tweaked if 1% proves the wrong value to use), with the current small data set, incluing the 'all' option is the only realistic way of having much of a chance of results when faced with a random set of ISBNs.

Possible Developments


The Read to Learn application is a demonstration prototype, and the code has not been robustly debugged or tested. Error handling is rudimentry.

Read to Learn is currently quite slow. The cause of this has not been tracked down, but it is expected that it could be made much faster.

Due to the fact that the MOSAIC data set only has information from the University of Huddersfield, the likelihood of matching any particular ISBN is relatively low. It is expected that if the MOSAIC dataset were to grow, matching would become more likely, making the service much more useful.