I blogged about my visit to Porto on Monday and I’m relaying appeals for help in two areas:
(a) One group wants to create an interactive map where they mashup immunogenetics data (genetic markers in human populations). The emphasis is on Europe. They are thinking of using Google maps, but I’m wondering whether Open Street Map is at a stage where it could create a totally open solution. I am gueesing that they wish to overlay maps with dots or coloured regions (so obviously natural administrative boundaries in Europe would be useful.)
Any help, including examples of where this technology has been used for whatever purpose would be useful.
(b) Another group is interested in publishing datasets in a data journal. They wish to make sure that the data are published at the same time as the study is published conventionally. They believe that the community would be interested in a journal which was peer-reviewed but primarily on data and metadata quality and that this would be “one data set per paper”.
This is close to the model of Acta Crystallographica E where each paper is a single crystal structure and where the data quality is the primary concern – chemical interest is not a determining factor.
I believe that the data journal is coming and that this type of activity should be strongly encouraged. It should be very cost-effective and could lead to new metrics for scientific endeavour.
OSM (OpenStreetMap) can be used. See info on the API at http://developers.cloudmade.com/projects .
Obama (http://www.whitehouse.gov/change/) and Tim (http://wiki.openstreetmap.org/images/f/f8/Tim_Berners-Lee_in_hi_viz_vest.jpg) are already on board.
I came across an example of an article that embedded interactive charts and data feeds the other day at: http://blog.fortiusone.com/2009/05/14/academia-20-what-would-a-fully-interactive-journal-article-look-like/
If nothing else, it provides an interesting strawman for discussing how to shape data rich articles, as well as how to store and publish the data that they contain in an accessible and open way.
I’ve recently started playing with Google spreadsheets as a possible home for data sets that can be represented within a simple spreadsheet. At a playful level, the API is easy enough to engage with (e.g. CSV output is easy to construct from a spreadsheet URL), and the provision of a query language that lets you treat the spreadsheet as a database could be quite a powerful feature (e.g. http://ouseful.wordpress.com/2009/05/18/using-google-spreadsheets-as-a-databace-with-the-google-visualisation-api-query-language/ )
More generally, I wonder whether we should all start looking more closely at the original definition of the HTML table element as a lightweight and easy to understand format for sharing data tables? (The YUI framework and Google visualisation API make it relatively easy to then progressivley enhance the tabulated data with visual representations.)
On the question of Google Maps vs OpenStreetmap, it may be worth taking a peek at mapstraction, which provides an abstraction layer above many mapping solutions. (I haven’t played with it yet but I v mindful that I should…) They just opened up a new sandbox too, as reported here: http://radar.oreilly.com/2009/05/mapstractionapi-sandbox-for-tr.html
For shiny shiny 3D mapping, Thematic Mapping just released a new API: http://blog.thematicmapping.org/2009/05/introducing-thematic-mapping-api.html
And of course, Google are making a play to own geo-data via their own ‘standards’: http://googlegeodevelopers.blogspot.com/2009/05/introducing-google-maps-data-api-in.html
@tony – many thanks. I’ll point them to this.
Crystallography is in a leading position on the way to the data journal, as Peter has suggested.
Robotics is another discipline where advances are occurring. See http://bit.ly/4xuoD for a recent announcement of “data papers” in the International Journal of Robotics Research.