On 04.06.2015 22:00, Scott MacLeod wrote:
Thanks, Markus and George,
I could only see a few controls in the upper left corner in Chrome in most of the links you both shared, except the video demo, George, did come through.
You need to wait until the loading is finished before you can see the map (the loading activity is indicated on the left when you open a page). Especially the "items" dataset that many links point to is pretty big. There are more than 2 million items displayed and filtered dynamically in this dataset, so the UI is not super-smooth. But then again it allows you to view more than 10% of Wikidata's items on a single screen :-)
Regards,
Markus
This is an exciting CC data which I hope will articulate with something like Google Earth, for example, and for a variety of STEM research foci, and in many languages.
Thank you, Scott
On Thu, Jun 4, 2015 at 8:51 AM, Markus Krötzsch <markus@semantic-mediawiki.org mailto:markus@semantic-mediawiki.org> wrote:
Hey, Thanks, Georg, that's really interesting. The added feature to inspect each element to see the items there is really useful. It's great that it works at high zooms now with the map in the background: I have spent some time exploring my own vicinity for people born there. As before, it works most smoothly for me on Chrome. You can see so many things in these maps depending on how you look at them. Here are some views that I found very interesting. If you click the links, it will always reload the data, so I am also describing how to make the settings manually. (1) "Popular places": the map of things that have articles on at least 20 Wikimedia sites: http://wwwpub.zih.tu-dresden.de/~s5219191/vizidata/#d=0&m=items&l=en&f=1&e=20_336&c=false&g=0.5&h=1.2&o=1&p=3&x=23.37890625&y=28.998531814051795&z=3 Item data set with interval narrowed down to 20-336. You can narrow this down further by sliding the left boundary of the interval at the top towards the right. The 1004 most popular locations on Earth are those with at least 96 sitelinks (as fewer items are returned, it is a good idea to increase the Size setting under Map Settings to see them more clearly). If you look at the things with more than 32 sitelinks, Italy somehow stands out among the rest of the world in terms of coverage. Even small villages there seem to be covered in many projects (why?): http://wwwpub.zih.tu-dresden.de/~s5219191/vizidata/#d=0&m=items&l=en&f=1&e=32_336&c=false&g=0.8&h=1.2&o=1&p=3&x=27.509765625&y=41.21172151054787&z=4 (2) Orphaned items in Wikidata: http://wwwpub.zih.tu-dresden.de/~s5219191/vizidata/#d=0&m=items&l=en&f=1&e=0_0&c=false&g=0.5&h=1.2&o=1&p=3&x=23.37890625&y=28.998531814051795&z=3 Obtained by decreasing the interval to "0-0" in the items view. This one has really interesting patterns ... you can see which countries have imported larger sets of items that are not from Wikipedia (having the Netherlands inside your map view slows down the browser considerably ;-). But it's also interesting to zoom in to some areas with few orphans to see what they are (spam? remains of deleted pages? something added by single users for some reason?). If you zoom all the way in to Amsterdam with grid size 0.2 you can recognise the structure of the channels: http://wwwpub.zih.tu-dresden.de/~s5219191/vizidata/#d=0&m=items&l=en&f=1&e=0_0&c=false&g=0.2&h=1.4&o=0.9&p=3&x=5.040321350097656&y=52.3465610683968&z=12 (3) People born between year 2BCE and 31CE (births dataset with interval set appropriately and increased grid size so you can recognise the few remaining points). http://wwwpub.zih.tu-dresden.de/~s5219191/vizidata/#d=1&m=humans&l=en&f=1&e=-1_31&c=false&g=2.7&h=1.2&o=1&p=3&x=78.3984375&y=25.005972656239187&z=2 This is nice to find errors since there are very few people at this time, but you can easily enter such a date wrongly if you type a date as something like "February 15" ;-). (4) Mortality peaks around World War 1 and World War 2: http://wwwpub.zih.tu-dresden.de/~s5219191/vizidata/#d=0&m=humans&l=en&f=1&e=1700_2015&c=false&g=0.9&h=1.2&o=1&p=3&x=-92.021484375&y=39.41922073655956&z=5 "Deaths" dataset zoomed in at North America. You can clearly see the blue peaks of death around 1918 and 1945, but you can also see that these increased deaths happened elsewhere (not in North America): the green bars are fairly smooth. Zooming in to a smaller time period of 1736-1883 on the same map, you can see that North America also had a peak in fatalities around 1864 that did not affect the remaining world: http://wwwpub.zih.tu-dresden.de/~s5219191/vizidata/#d=0&m=humans&l=en&f=1&e=1736_1883&c=false&g=0.9&h=1.2&o=1&p=3&x=-92.021484375&y=39.41922073655956&z=5 Cheers, Markus On 04.06.2015 13:41, Georg Wild wrote: I'd like to once more draw your attention to ViziData [1]. The project has matured a bit since I first brought it up here a few months ago. Most notably there is now an underlying tile map providing some orientation better integration and interaction between map and timeline and the possibility to list items aggregated in a map cell. I created an overview page [2] for the Wikidata Visualization Challenge, which can aid as quick reference for usage of the application (although it is pretty cluttered). The version linked on that page is my submission to the challenge and already outdated though, the most up to date version should usually be found on the link in this mail [1]. There are only 2 new datasets, one is all locations that have a statement about their populations, however that needs some more work and fine tuning. The other is any item that has a coordinate location statement, filterable by the number of interwiki links. It shows a pretty good coverage of the world, bring some patience though because the 60mb json can take some time depending on the connection. Also, slower hardware might get a bit exhausted going through the ~2.1 Million items. I guess one can imagine various use cases for the application. For instance, here is the people who died in the Netherlands in 2015 in Vizidata [3]. Be aware however, that those haven't necessarily been Dutch and also it doesn't include Dutch people who died somewhere else. Comparing with Gerards list [4] there seem to be quite some people missing. That's mainly because items that don't have both a date of death and a place of death statement are dropped from the datasets. Maybe it would be an option to include country of citizenship as a fallback but it could lead to slightly incorrect representation of the data. I hope this tool can provide some interesting insights and help to illustrate where there is most room for improvements in our data base. Feel free to comment if you have questions or proposals, or maybe an idea for a dataset that you would like to see in ViziData. Georg [1] http://wwwpub.zih.tu-dresden.de/~s5219191/vizidata/ [2] http://wwwpub.zih.tu-dresden.de/~s5219191/vizichallenge/ [3] http://wwwpub.zih.tu-dresden.de/~s5219191/vizidata/#d=0&m=humans&l=en&f=1&e=2015_2015&c=false&g=1.6&h=1.6&o=1&p=3&x=5.75958251953125&y=52.52624809700062&z=8 [4] https://www.wikidata.org/wiki/User:Jura1/Recent_deaths_in_the_Netherlands _______________________________________________ Wikidata mailing list Wikidata@lists.wikimedia.org <mailto:Wikidata@lists.wikimedia.org> https://lists.wikimedia.org/mailman/listinfo/wikidata _______________________________________________ Wikidata mailing list Wikidata@lists.wikimedia.org <mailto:Wikidata@lists.wikimedia.org> https://lists.wikimedia.org/mailman/listinfo/wikidata--
- Scott MacLeod - Founder & President
- http://worlduniversityandschool.org
- 415 480 4577
- PO Box 442, (86 Ridgecrest Road), Canyon, CA 94516
- World University and School - like Wikipedia with best STEM-centric
OpenCourseWare - incorporated as a nonprofit university and school in California, and is a U.S. 501 (c) (3) tax-exempt educational organization, both effective April 2010.
World University and School is sending you this because of your interest in free, online, higher education. If you don't want to receive these, please reply with 'unsubscribe' in the body of the email, leaving the subject line intact. Thank you.
Wikidata mailing list Wikidata@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikidata