Over the last twelve months i invested about two full and good weeks of time to get to know wikidata and bring the Wikidata Topicmaps UI (WTUI) live. The last development days were mostly spent on getting images and graphics related to wikidata items directly displayed the Webclient. I could do this through additionally connecting the Commons Media API by Magnus Manske. Today was the day to upgrade the service about the developments of the second half of 2014 and now there is this ready-to use web-app i wanted to tell you about. It is hosted at Wikimedia Labs, provided under the Terms of Service of the Wikimedia Labs Projects and currently maintained by (a researching/volunteering) me. If you don’t want to rely on me you can simply get your personal installation running after downloading the DeepaMehta 4.4 Standard Distribution and the dm44-wikidata search plugin, both are to find at http://downloads.deepamehta.de. The application works well in most browser though there is currently no support for Internet Explorer users. Presenting your Topicmaps to an audience requires either an internet connection or a personal installation on your PC, but “readers” of your maps do not need an account.
The look & feel of the WTUI is still a bit rugged at the moment but as i wrote in the post on “situational apps”, its a good and simple enough to be used and to be developed further by or with others. If we find an interested user base, why not get together and bring this a few steps further?
As a kind of foreword i want to link here to a post in which Javiera Atenas, Leo Havemann and Ernesto Priego wrote about make use open data sets (like wikidata) in educational contexts. As they note, from their perspective “research is currently lacking on open data as educational resources, and about how students can develop critical skills by using open datasets.“. To support and connect researchers and users of both communities, deepamehta4 and wikidata was the main motivation for this project to start. Now using the Wikidata Topicmaps UI (built into deepamehta4) will at some point very well turn everyone of us (deeamehta4) into a critical and contributing user of wikidata.org.
The stage is set and we will see who steps up. In the future, building on the results presented here i would like to improve on the current relationship between being an editor of wikidata and being a user of deepamehta4. At best these two “roles” will become one.
The data at hand here reaches over languages, linguistics, culture, politics to openness and with this web-app we can visualize items from various domains in their complexity. Furthermore we can also take a deep look under the hood and see how we (as the wikidata community) structure our most common information. Intriguing thoughts, no? At the end of this post you’ll find my first screen-recording/video in years, but i have to disappoint you, it comes without sound.
Let’s start with having a closer look on what we have right now
1 To be able to create, store (and thus develop) personal Topicmaps you need to register an account and log in with your desired username. As mentioned, this is necessary to provide you the opportunity to create your personal maps as well as to comply with the ToS as we are an official Wikimedia Labs project.
2 After you got your personal username you log in and will be automatically redirected to the here. In the Topicmaps UI you can select between all existing “Topicmaps” through using the corresponding drop down menu in the upper “Toolbar” and explore what pubic maps other users have already composed. In the “Topicmaps” menu you will also find an entry called “New Topicmap”, click that for starting to create your own view from a blank screen.
In the following picture you see a map on “Countries & Political systems” arranged after surfing about one hour along specific relations of items describing “states”. As these topics represent regions of the world, it has become quite a different image compared to a geographical picture, where Germany would rather be further away from Argentina then in this visualization (in which both share the attribute of being a “federal republic”).
3 Constructing a map is done through using the two main commands provided by this wikidata plugin and the standard means of interaction provided by the deepamehta 4 webclient:
- the “Wikidata Search” mode in the DeepaMehta 4 Webclient Toolbar (shown in the following image) and
- the “Import claims” command provided by all “Wikidata Search Result” items
This search mode lets you choose among many language in which (a) is searched for the text you enter and (b) the language in which you want to import all the contents. After pressing “Search” the search request is sent to the Wikidata API and if successful, the application will present you a list of search results along with some context information. Clicking through the search results will populate your canvas and lead you to the next possible command.
4 Importing “statements” and all navigating to all related items
In wikidata items or topics are described through so called “statements” (or “claims”) which every user can author. These statements follow a strict rule and structure which you will be able to explore and get to know after importing all the statements for any topic of interest. If you find an item with no further details on the right side of your screen (within the so called “Page Panel”), you first need to trigger the “Import claims” command at the bottom of the page (or from the context menu of the item).
This feature must be run at least once on every topic you want to study and its functionality is to copy all labels, descriptions and aliases which are currently stored for this item on wikidata.org into your personal database. Also this command imports short infos on all items related to the, e.g. Country at hand. Note that this command may take as much as 15-30 seconds if you have items about there is a lot of information in wikidata, so you might want to prepare this before presenting your map towards an audience.
The standard interaction methods of controlling what is displayed on the so called “Canvas” is using drag & drop, panning as well as the commands available in the context menu (via right clicking) like “Associate” and “Hide”. Surely you can furthermore use all other standard features of deepamehta4 to “Create” and relate “Notes” or other topics to your wikidata entities.
I tried to demonstrate all these points in the following, silent screencast of me clicking around in Wikidata Topicmaps.
Useful features in Topicmaps for educational purposes
In general, this UI allows (and even demands manual) free placement of topics on an unlimited canvas. But remember, all views are stored independently of the contents (items), which means that you can reveal every other topic in your database through navigation or search in any other Topicmap at any time.
When using wikidata topicmaps in a classroom the following features might be noteworthy
- Stable re-positioning after Hide / Reveal: Once a specific item was revealed in a Topicmap its position is persisted/stored within the map, which means when the topic gets hidden and revealed one can be sure of the position where it gets revealed. This allows you to customize the construction and revelation of topics in front of pupils and/or audience in advance of a session (e.g. overlapping edges or texts in the map which would make it harder to read)
- The systematic colouring of topics is possible through using the “Set Color” command available on each topic (context menu).Tthis way you can further chunk or group topics, as i did with countries in the example above.
- Background images: You can upload a bitmap image file and make the webclient render it as the background of your network of topics. This may be fancy and allows you to provide custom semantics or mood images on the background of the to be discussed data.
Some details and the request for user feedback
The first idea behind this project was to enable wikidata users and editors to quickly create and maintain personal perspectives on wikidata as a vastly developing strucured data set. As you will find out using this service, despite some lacks in usability and comfort, this goal could be regarded as achieved by now. Nonetheless it is important for me to improve on what we have now, bring you and many others in here and develop shared ideas, especially because this app-development was not directly built (with the aim) to support teachers, pupils, facilitators or students. So please, i am happy to support more research and development here and always happy to help make other great things possible. Therefore i am looking forward to you sending me your user feedback directly to malte@deepamehta.org or leaving me a comment here.
Notes on this open source development
To conform with the Wikimedia Labs Project guidelines and to bring to you the best of the underlying software platform i alsoe needed to code a plugin which brings support for self registration of user accounts to the dm/x platform. The code of this sign-up module for dm/x is hosted here and a downloadable release is available on http://downloads.deepamehta.de.
Here are the latest details on the development of the wikidata-search 0.4 source code, released at December 27th in 2014.
- Integration of Commons Media Files into “Import claims” command
- Added a couple of languages to choose from for searching
- Improve the rendering of wikidata properties in the page panel
- Forked the branch on wdtk integration into a new wikidata plugin: dm4-wikidata-toolkit (on which i plan to report more next week here)
- Upgraded this Wikidata Search plugin to be compatible with dm/x 4.4.
Upgrading to the new, more collaborative DeepaMehta 4.5 is pending because that is something i can’t just start and finish today or within the next week.
Nonetheless happy searching and clicking and i hope you sign-up!, give it try and send me some feedback!