This is an old revision of the document!
OpenLetten
This project is less about “data” than the others, and more about “open”. Letten is a place at the Limmat river right behind the ZHdK building (where MakeOpenData is hosted) in Zürich, where lots of people like to swim.
Photo by Yago Veith www.yago1.com used under (CC BY-NC-ND 2.0)
Daily temperature measurements taken at this exact location are readily available on the Web, but they are published in a hard to parse PDF document oriented towards “scientific” readers. This projects shows what is necessary to make use of the same data to answer the simple question “Should I go swimming today”. Different visualizations are shown as a proof of concept. Finally a mashup connects the temperature to an LED to get a basic, physical “ambient notification”.
The aim of OpenLetten is to show how hard it is to parse data if the publisher choses to only support the human readable data format (here PDF ), and how easy it becomes to build novel visualisations and mashups once the data is also published in an open, machine readable data format (e.g. CSV or RSS ).
Data
First, the PDF data is parsed by HwPdfToPachube.exe, a rather complex C# program given the simple task of extracting the current temperature. IT took me 4 hours to write it. The data is then published to Pachube, a sensor platform based on open data formats. There might of course be other such platforms that could be used.
- http://www.hw.zh.ch/hochwasser/mac/at/0851.PDF (Source data, PDF )
- https://pachube.com/feeds/36745 (“Open” data, available as XML, JSON, CSV feed)
Visualization
Once the data is available in an open format it's quite easy to consume and display it in a way that suits the consumer rather than the publisher of the data (the visualizations below contain about 3 lines of code to get the data, compared to the 4 hours above). While the original PDF contains a lot more information, the current temperature is what's most interesting if you're about to decide if you should go for a swim. There is no way for the government to cater to such needs as they might be very special or of concern to a small minority only. Data in an open format empowers such groups to take action themselves and use the data they paid for with tax money in a way that serves them well.
- http://www.tamberg.org/makeopendata/2011/pachudial.html (PachuDial visualization, Flash, animated, live)
- http://www.tamberg.org/makeopendata/2011/jquery.html (Simple visualization, JSONP, live)
- http://www.tamberg.org/makeopendata/2011/googlecharts.html (Google Chart visualization including comfort zone, JSONP, live)
- http://try.yaler.net/~arduino/led (Arduino LED with a REST interface, LED color set by HTTP PUT request from LettenLedMashup.exe, sorry for the ads, 504 if offline)
UPDATE, 29.03.2014
Here's a picture of an OpenLetten widget developed during the Samsung Smart TV Hackathon 2014.
Team
Source
- https://bitbucket.org/tamberg/samsungsmarttvhack (29.03.2014)
Links
- http://sourceforge.net/projects/itextsharp/files/ ( PDF to text C# library)
- http://www.paravan.ch/geocaching/#tools (CH-1903 to WGS-84 coordinate converter)
- http://code.google.com/apis/chart/ (Image and Javascript API for visualisation)
- http://en.wikipedia.org/wiki/JSONP (Cross site scriptable JSON API convention)
- https://pachube.com/ (Sensor data platform based on open standards)
- http://www.instructables.com/id/Arduino-Web-LED/ (How-to, including source code)
- http://yaler.org/ (Access to LED Web service running on Arduino; Disclosure: I'm a founder of Yaler)
- http://www.thingiverse.com/thing:148319 (09.09.2013)
Back to project overview: http://makeopendata.ch/doku.php?do=search&id=@project