Yesterday we released an updated website for publishing UK and European product recall notices. The changes we’ve made at RecalledProducts.org include a refreshed look, enhanced statistics and an improved recalls importer.
The biggest change you’ll notice is the design of the site. It’s more of an evolution of the design than a brand new look – we’ve kept the same structure, but given it a bit more of a “CF Labs” style (similar to StayPrivate.org and Community Post Offices).
We’ve increased the width of the site which has allowed us to improve the layout of the product recall notices. We’ll be making better use of this extra space on other parts of the site over the coming months.
We’ve also spent some time working on the recall statistics we generate – in some cases, we felt they weren’t particularly clear. For example, the chart showing the top dangers in recall notices included “chemical” near the top, but further down it also included “chemical and damage to sight”, “chemical and drowning” and “injuries, choking and chemical”.
This meant that some of the figures weren’t fully representative of the real numbers. As a result, we’ve completely rebuilt the statistics engine to take account of this kind of thing. This has also enabled us to output the data in CSV and JSON formats for those that want it. As always, these statistics are available under the Open Government Licence.
The other big change we’ve made is to the scraper we use to import recall notices from the European Union RAPEX system. Occasionally odd characters crept onto the notices where certain special characters hadn’t been encoded properly by the source. We’ve updated the scraper so it handles this encoding on their behalf prior to importing to reduce the likelyhood of this happening in future. To improve the existing data on the site, we’ve gone back and re-imported the latest 300 recalls (everything from January 2011 and most of December 2010).
Occasionally RAPEX make changes to notices that have been published such as fixing errors, adding additional information or removing notices that have been cancelled. While we were picking up and replicating these changes before, our updated scraper allows us to detect these in much greater detail.
To ensure users can see an accurate history of each recall notice, we’ve now included a column on the notice’s page that will list any changes made. Where a text change has taken place, we’ll include the original text for comparison. You can see an example of this on one of the recalls we re-imported yesterday.
We’ve got more improvements planned over the next few months, so keep an eye out for updates on this blog.