Wikipedia

User:Aude/GIS

< User:Aude
GIS

One of the greatest shortcomings I see with Wikipedia and Wikimedia projects is with our use of maps and geospatial data. I have seen ideas here and there on this mailing list and other places, but it's a big issue to work through.

Map requirements

edit

Articles found using Special:Random that could use maps.

Need maps

edit

Locator maps

edit

Other maps

edit

Have maps

edit
  • Ilava District in Slovakia - this map is a .png, it would be difficult for someone else to make use of the information here to make new or different map of Slovakia. If we had a GIS file with this information, it would be helpful.
  • Magyaratád, a village in Hungary - this has a .png map, showing the location of Somogy (the county where this village is located). Again, it would be helpful to have these sub-national boundaries as a GIS file, so they can be reused for various purposes. For this village, a locator dot map would be helpful.
  • Mona Township, Ford County, Illinois - this is a .png map. Shows the location within the county, but would be better if the map also showed the location of the county within the state to help orient readers. Since this is a place in the U.S., it should be possible to obtain public domain GIS files from the Census Bureau or elsewhere that would include these boundaries.
  • Pennsylvania's 8th congressional district - this is a .gif map, which came from the National Atlas website, which is a GIS-driven mapping site operated by the USGS.
  • Bad Suderode, a municipality in Germany - has a dot locator map, which is based on Image:Karte Deutschland.png. Would help to have a GIS file with these subnational boundaries.
  • Kilchberg, Zurich, a municipality in Switzerland - has a dot locator map, with a .png base map.
  • Gallitzin State Forest - located in Pennsylvania, has a dot locator map with a map of the entire U.S. as the base map. This is the sort of map that could really use the ability to zoom in and see something more than a dot on a U.S. map.
  • St. Augustine Alligator Farm Zoological Park - the dot locator map, with a U.S. map as the base, is entirely inappropriate here. The dot is misplaced and shows a location in Georgia, not Florida. This needs a map zoomed in more.

Core topics

edit

Creating locator maps

edit

Our approach for locator and reference maps in articles is very rudimentary and ad-hoc. I have background in geography/GIS, and ability to use state-of-the-art software such as ArcGIS (http://www.esri.com) that is generally beyond the means of consumers, and knowledge/skills on how to use it to make maps for Wikipedia. (though, the company makes its software available at educational prices and sometimes donates software).

There seems to be a zillion maps that could be made - locator maps, thematic maps (e.g - choropleth maps that show census variables), etc. It takes significant time, though to make them, and do them well. For example, I created a map for Shoshone National Forest, but to do more of these maps in a systematic, standard all other protected areas in the U.S is a significant undertaking. Then there are places in Canada, and around the world that would need maps. It would take a while, but can be done.

Hardware requirements

edit

Working with GIS also requires large amounts of disk space to store and work with the data (especially if you bring in satellite/aerial imagery), and decent hardware. I have ~terabyte of hard-drive space, spread out among numerous hard drives. Yet, I am running out of hard drive space, and go through disk space quickly. It's a hassle to keep my GIS data all backed up properly. With data stored on various local hard drives, it's difficult to share data and work with it. When I run out of space, sometimes I delete GIS data that I worked with, such as Census tiger data or satellite imagery (I can always download it again later). I now have a dedicated web server, which I intend to use to store GIS data and make available through a web-mapping system, and archive map data that I have used to make maps. Still, disk space and bandwidth will be an issue.

Static maps

edit

I realize for making printed versions of Wikipedia, having static maps is important. Creating them is time intensive, may require specialized software (can be expensive), specialized knowledge and skills, and powerful hardware to handle the large GIS data inputs. Static maps for online use are less than ideal, in my honest opinion. Depending on the topic, static maps can become outdated.

Despite this possibility (with significant effort), I'm not entirely sure having only static, non-interactive maps (quickly can become outdated) is the way to go. I haven't come up yet with an ideal way to solve this issue. I could probably write scripts on the GIS-end of things to help automate it, and a bot on the wiki-end to ease the process. I have also played around with the MediaWiki source code and been learning about writing extensions. If we could come up with some possible ideas for providing better, (more standard?) maps.

Ultimately, it may be useful to have a map generator/machine application, which retains source GIS files so that others can use and re-use them. As a wiki, anyone can contribute map data. Possibly, like Wikimedia commons which makes image thumbnails and stores them, perhaps a map-maker application could make static maps and store/cache them so they are ready to use. When data or information on a map changes, the it can be uploaded to the map application and new static maps created, in more of an automated or semi-automated manner. In the U.S., with National Atlas, such a map generator exists to some extent. But, it's U.S. only.

GIS data repositories

edit
Main page: User:Aude/GIS data

And a great deal of geospatial data in GIS formats:

There is also a great deal of public domain satellite/aerial imagery.

With these formats, specialized software and capabilities is often required to make use of this data.

Offline maps

edit

There is a great deal of maps and data out there that is in print format. It would be fantastic to be able to use resources such as the "collection of rare maps of Africa, dating from 1530 to 1915" at Northwestern University that's been mentioned. In order to make best use of them, it may be helpful to have maps georeferenced (geoTIFF is the format for this). In addition to this collection at Northwestern, check out http://www.lib.utexas.edu/maps/faq.html#3.html (see "Where are the rest of your maps?") And, the map library at Penn State (http://www.libraries.psu.edu/maps/) is quite extensive. I know some folks there, might be possible to work with them on something if we could come up with a specific idea.

Other sources

edit

Aside from public domain, U.S. federal government sources, there is a wealth of geographic data held elsewhere, such as local governments (street centerlines, neighborhoods, etc.), or other countries (e.g. Canadian federal government) that is made available to the public. But, it's not explicitly public domain.

edit

Geographic data is increasingly being distributed with licensing, rather than outright sale or free distribution of data products, for various reasons including to limit potential liability with explicit language.[1] This is the case with data made available by the Government of Canada, which puts liability clauses in the license as well as retaining Crown Copyright, but otherwise essentially now allows free access and use of GIS data. It is unclear if one takes some of this data, manipulates it significantly, combines it with other data, and makes a map with it, what the copyright status of the map is. It might be one's own work, though it's always important to cite and credit your data sources. Perhaps, if we had a relationship with some of these data holders, maybe some would more explicitly agree for their data to be used.

In the U.S. government, data is traditionally in the public domain, but increasingly the government is contracting out for projects and labor and relying on the private sector more. In the case of high resolution satellite imagery, instead of the government building its own satellites and competing with private sector companies such as DigitalGlobe, Space Imaging, and GeoEye (formerly ORBIMAGE), the government buys imagery from those companies. In 2003, the National Geospatial-Intelligence Agency (NGA) established ClearView programs, which permit acquisition of high resolution commercial imagery by the agency. The ClearView license also "permits sharing imagery with all potential mission partners", meaning that other government agencies can get no-cost access to the data, but not the public.[2]

United Nations

edit

The United Nations has a treasure-trove of GIS data, such as the GEO Data Portal, which includes 450+ variables at national and subnational level. Much of the data covers years going back to the 1960s for topics such as infant mortality, arable land, etc. However, there is a copyright notice at the bottom of pages on this site, and the user agreement specifically states "Use of the "GEO Data Portal" for commercial purposes is strictly prohibited." Thus, the data is not compatible with Wikipedia. Complicating things is that often data is provided to the UN from other organizations, including from national governments, who themselves have their own copyright restrictions.

Non-spatial data

edit

Other data, such as U.S. census data (going back in time) is available and can be "joined" to geographic data, and mapped out.

GIS software

edit
Main page: User:Aude/GIS software

Open source GIS

edit

Commercial software

edit

Previous/Current efforts

edit

Atlas portal

edit

I see that we have an "atlas" portal on commons which looks great: http://commons.wikimedia.org/wiki/Atlas

But, I find it lacking in content, with maps (especially of places outside the U.S.) to be overly general: http://commons.wikimedia.org/wiki/Atlas_of_India

Geocoordinates in Wikipedia

edit

I'm also aware of the effort to add geocoordinates to Wikipedia articles, and that Google Earth is now incorporating these.

Having geocoordinate points is not always suitable. A place such as a city, state/province, country, or features such as a river are area (or sometimes linear) features. Sometimes, geographic entities or phenomena are best represented with raster, instead of vector formats.

What would be ideal is to have GML file(s) associated with an article; Depending on the article/place, the geographic feature might be a single point, line, or polygon.

Map/GIS extensions

edit

Other projects

edit

Integration

edit

Because of the vast possibilities, and significant undertaking it is to work with geospatial data and possibly integrate to Wikipedia, we could use some resources such as grant $. Alone, the hardware needed to be able to accept geospatial data on Commons or somewhere might be significant.

Maybe we could brainstorm and have discussion to learn what kind of maps or mapping capability that people would like to see with Wikipedia? Maybe then a focused, pilot project on a specific set of places or articles.

I think it would be excellent if we could:

  1. Help get offline, printed maps digitized and georeferenced into a useful digital format (e.g., geotiff is commonly used for this)
  2. It would also be excellent if Wikimedia commons (or other Wikimedia project) could accept geospatial data/files in vector format (GML is becoming the standard - http://www.opengis.net/gml/), svg can be used to display such data.
  3. Being able to integrated maps, imagery, GIS data, Wikipedia geocodes, and come up with a way to get at information in a geographic way.

References

edit
  1. ^ Committee on Licensing Geographic Data and Services Staff (2004). Licensing Geographic Data And Services. National Research Council. pp. p. 3. ISBN 0309092671. {{cite book}}: |pages= has extra text (help)
  2. ^ National Geospatial-Intelligence Agency (January 17, 2003). "NIMA Partners with Remote Sensing Industry". Directions Magazine.

Further reading

edit
edit
Retrieved from "https://en.wikipedia.org/w/index.php?title=User:Aude/GIS&oldid=786790835"