Section I: Finding Data
Throughout this tutorial you've been provided with data that you've used to work through various exercises. Once you're working on your own projects, you'll need to find or create the data you need. There is a lot of free GIS data available on the web, created by various government agencies, academic and non-profit organizations, and private companies. You can try a search engine or look at an academic map / GIS library website for a list of helpful links (a list of suggestions is included in the following section). To be strategic about your search, it helps to understand who creates and provides the data:
- Global / international: Look at supra-national agencies, like the United Nations (in particular, the UN's Environment Programme has a good site) or academic / non-profit organizations who have enhanced and updated public domain data such as the Global Administrative Areas (GADM) site, the DIVA GIS data page, and the Natural Earth project. If you need satellite imagery the best sites to visit are the USGS and NASA.
- Country level: In some cases you'll want to visit a few of the international sites, like DIVA GIS and Natural Earth, to get basic country-level datasets like state or provincial boundaries. But in many instances you may want to visit a mapping agency website or data depository for the specific country you're interested in; you'll find more country specific layers and they will be processed in a way that is readily compatible for mapping attribute data from that country. Most countries have one or two agencies that will provide the bulk of the country's GIS data - a statistical agency responsible for the census, or a mapping agency responsible for surveying. In the US you could go directly to the US Census Bureau or the USGS to download data, or you could visit the central data.gov repository. In Canada, you could visit Statistics Canada directly or visit the Geogratis repository. Some countries may provide one central source (Australia), whereas other countries may provide limited or no data, via a website that may or may not be in English.
- State / Provincial: You may be able to visit a country level source, like the US Census Bureau to get state, county, or zip code boundaries for the entire state, or you can visit a state level agency to get more specialized datasets for that state. Some states will have state government portals where you can access all data for a state, others may cooperate with a college or university located in that state to provide data via the university's portal. In addition to centralized portals, individual departments or agencies may also provide data directly; road and transportation layers may be provided by a state department of transportation or may be provided through the state's central portal. State agencies are also the most likely source for aerial photography.
- County / City / Local: Local governments may have portals where they provide administrative boundaries, transportation data, and real estate or tax parcels, and datasets that would be of local interest (such as neighborhood boundaries that may not be formally defined elsewhere). You can also look at the geography one step above (state level) to see if data is available for the local area.
- Gazetteers and Geocoding: if you can't find an existing GIS dataset, you can always try to create one from an online gazetteer that provides latitude and longitude coordinates for point-based features; the USGS has a US level gazetteer, while the NGA has an international gazetteer. Do you have a list of addresses but no coordinates? Try uploading them to a free geocoding service like the one at USC GIS Research Laboratory, which will translate your addresses into coordinates.
- In some cases you may find university or non-profit sites that provide data within a specialized area of interest. While universities typically provide data for the geographic areas where they reside, there may be special labs or research groups that provide data beyond that area; the CIESN (Center for International Earth Science Information Network) site at Columbia University and the NHGIS (National Historic GIS) at the University of Minnesota are two examples.
Regardless of where you download your data, you'll want to examine the metadata for the layers. Metadata can be formally or informally described on the website where you downloaded your files, in narrative documentation that is included with the files you downloaded, or in special XML files that accompany each of your GIS files. There are a few well-defined standards such as the FGDC and ISO 19139 that data creators use to document data, and include elements that explain who created the data, when it was last updated, what the file contains, what the intended purpose of the file is, if it was created for a specific optimal scale, the coordinate system and map projection it was created in, and copyright and use restrictions. You'll want to check the metadata to verify that the data is going to meet your needs and that you can use it for your intended purpose. For example, you wouldn't want to use a generalized boundary file if you're mapping at a large, local scale, and if you are going to use the data for a commercial purpose you need to verify that that's permitted. In any event, you should cite the source of your data in any maps, tables, or reports you create from it.
If you are looking for a particular GIS file and it's provided by several sources, which source should you use? For example, if we wanted census tracts for a particular city, we could download them from the city's GIS page, from a state-based site, from one of ESRI's pages, or from the Census Bureau itself, via the TIGER page or the generalized boundary page. To answer this question, you'll have to examine the download page, and even download the files to view them and their metadata. Here are some things to consider:
- How are the files packaged for download? Do I have to download them one place at a time, or could I get the entire area in one download?
- Who created the files originally? Is it better to go with the original source? Or has a secondary source added some value that makes their files more desirable?
- Can I trust the source? Is there metadata? How did they create the data?
- For vector files, are the layers generalized or not? What scale are they appropriate for?
- For vector files, are the polygons saved as single or multipart layers?
- For vector files, what attributes are available in the attribute table? Are there ID codes that I can readily use to join data? Are there place names that I can readily use as labels?
- For raster files, what is the resolution of the data? Is it appropriate for my intended use?
- What format is the file in? Is it a format I can use, or at least one that I can easily convert?
- Are there any copyright or use restrictions with the data?
Finally, remember that GIS data is often just one piece of the puzzle. It represents the geographic features, but if you need attributes to go with these features (demographic data, weather data, sales data, etc) you'll have to download this data from someplace else (or create it yourself) and process it to make it usable with your GIS data.
