User:RolandUnger/Happy Birthday, Wikidata!
Wikivoyage likes to wish a happy birthday to Wikidata. Wikidata turns five—a real long term on the web—and only a few months after Wikivoyage turned ten.
Wikidata helped Wikivoyage to become more attractive and useful. Thanks!
Only two years after the establishment of Wikivoyage we added on October 17, 2008 a new feature to that wiki: a location data base. It was our own data base to store interwiki links, breadcrumb trails, coordinates, sizes, heights, and so on. For instance it was possible to enter minimum, average, and maximum values for mountain heights including their precisions.
About five years ago Wikivoyage became a Wikimedia sister fond of traveling. Unfortunately, we had to forgo our location data base. In 2013 we learned about the new Wikidata sister project. Already on August 27, 2013 our first Lua module on the German Wikivoyage was brought in to action to present interwiki, intrawiki links and coordinates which were taken from Wikidata. At this time we already were sure that Wikidata could be a real substitute for our own data base.
In 2016 and 2017 there were several milestones of using data from Wikidata in Wikivoyage articles. This could be done only by extensive use of Lua modules. The infobox, named quickbar at Wikivoyage, is able to fetch all data like region, count of inhabitants, contact information of tourist offices, and for location maps. All data can be overwritten by local ones but differences between them and a link to edit Wikidata are presented in case of missing and different values, too. The most difficult question was how to get the name of the first-order administrative entity for all countries of the world?
Our listing template was a new challenge using Wikidata. The template presents all data of institutions like hotels, museums, and so on. About 40 parameters can be specified, and most of them like local and international names and addresses, coordinates, phone numbers and booking.com hotel ids can be fetched from Wikidata. Unfortunately the programming took not only one month but at least twelve. There were three main problems: (i) simplifying the instance of an institution (is an interchange station a railway station?), (ii) handling overseas territories, (iii) slashing the computing time the by a factor of 5 to 10. Otherwise, 200 templates within a single article would bring the Wikimedia software to its knees. Filling this template is supported by a special editor which—of course—presents all data which are already available at Wikidata.
There were about 3,500 Lua code lines to make the listing template working. We hope that SPARQL queries and the improvement of the Wikidata-Lua interface could help us in future to simplify code and reduce computing time.
A future step should be the transfer of data stored in Wikivoyage articles to Wikidata which needs a lot of manual support to ensure high data quality. But it will help at least small Wikivoyage communities to bring these data into their articles and to automatically update them. A few years ago, an idea was born to add a review forum for the institutions mentioned. But this can be done only with a data base like Wikidata storing all the institutions’ data.
Unfortunately Wikidata cannot help in all cases. Mass data like complex geographic polygons and shapes and climate data are stored now on Wikimedia Commons or OpenStreetMap. In case of geographical data stored as GeoJSON it is already possible to present them on maps in articles. And we hope that climate data can be shown in diagrams soon.
It’s much to do in future. But together we will succeed.