Common use of Formats Clause in Contracts

Formats. For every data set we chose from provided formats the one, which was best in terms of usability, openness and data quality. Finally we ended up with a list of 31 different formats, where 16 formats were only available in four or less data sets. A big share of open data is spatial, almost half of the data sets, which makes that WFS was the most common format with the share of 32%. The wide adoption rate of WFS is possible partially thanks to Inspire directive, which defines requirements on what and how spatial data should be published (The European Parliament and the Council, 2007). After WFS follows CSV, with a 26% share. In practice CSV is also an output format of WFS-standard meaning that over 60% of all data sets are accessible in CSV format, making it at the moment the most commonly available open data format. But as JSON, WFS and other formats provide enhancements to CSV; therefore they are valued over it. A big share of 86.5% of the open data sets is in easily machine readable formats. Five most common formats, which are not easily machine-readable, are: XLS, ZIP, ODS, PDF and HTML. Although all of them can be read programmatically, practical implementation of reader is complicated and in any changes of structure the read process will probably fail. One easily doable movement towards better rated open data is to publish all Excel and ODS based data sets also as CSV files. The division of different data formats in cities’ open data portals is illustrated in Table 4. WFS 198 32,2 % yes CSV 173 28,1 % yes API 35 5,7 % yes WMS 27 4,4 % yes Excel 26 4,2 % no JSON 26 4,2 % yes SPARQL 23 3,7 % yes ZIP 16 2,6 % no ODS 11 1,8 % no PDF 11 1,8 % no SOLR 10 1,6 % yes HTML 9 1,5 % no ZML 8 1,3 % yes RSS 5 0,8 % yes WMTS 5 0,8 % yes Others 32 5,2 %