Design for the final resolution, size, and viewing environments in which readers will use your maps.OPTIONS FOR GETTING YOUR MAP OUT OF A SPECIALIZED GIS environment and onto people’s screens or desktops are ever increasing. The task requires you to plan for different display resolutions, both finer and coarser than your GIS interface presents. Combinations of resolution and map size also affect your export choices. Basic export formats fall into two categories—raster formats (Joint Photographic Experts graphic [JPEG], tagged image file format [TIFF], portable network graphics [PNG]) and vector formats (Adobe Illustrator [AI], portable document format [PDF], scalable vector graphic [SVG]). Vector formats can include raster layers, and PNG formats can handle transparency. Some of the vector exports keep words intact as searchable wholes (AI), while others break them into individual letters or segments of words (PDF). These considerations affect the ease of additional editing beyond GIS. Resolution, scale, and file format come together as you decide how to tile or cache maps as web map services. Your knowledge of these issues ensures your best map designs are not diminished in the last stages of dissemination.
Many maps are intellectual property that are protected by copyright laws. Ask permission to use the content of others, or choose content that has explicit public licenses, such as those recommended by the Creative Commons. Credit the sources of data, maps, and images used for your maps in published and online products. A map or dataset that is freely available is not necessarily free for use in a commercial product or on your own websites. You should also expect others to honor your creative mapping efforts with attribution and by requesting permission to use them.
The following list includes the essentials of sharing and publishing well-designed maps:
Design for the final resolution, size, and viewing environments in which readers will use your maps.
Choose raster or vector export formats that best present your work, with attention to label and transparency quality.
List your sources for both data and graphic elements on a map, even when they are in the public domain.
Follow licensing restrictions for data and maps you use and list your copyright or public-use licenses clearly so others know how to use your work.
Choosing how to present a map is part of the design process. Maps are designed for multiple and varied contexts. Each context will be best served by a different map design. Consider a few places we commonly find maps:
full computer screen viewed at the reader’s desk
projected display presented to hundreds of people at once
color laser prints distributed to a working group
black-and-white print for a report that concerned citizens will hand out as photocopies
plotted poster pinned up at a planning meeting for viewing from across the room
page in a glossy magazine or book that is professionally printed on an offset press
huge backdrop at a trade show
supporting information in a documentary television show
black-and-white fax to an emergency response team
small Retina display on a mobile device used during route planning
part of an online interface for web-based data dissemination
Each of these modes of display constrains how a map can be made and what it can contain while still being legible. Rather than complain about (or worse, ignore) these constraints, your job as the mapmaker is to use good design to master them. Many of us have attended a talk where the presenter declares that the projector is at fault for the illegibility of the maps. Wrong. The error is made by the presenter who borrowed a design suited for another context or by the map designer who did not account for the final display constraints. If you need a map in a projected presentation, redesign it with bolder color differences, larger type, and simpler lines to be sure that the main messages will hold up at coarse screen resolution, bleached by the projector and the room lights, and viewed from a distance. If that map is printed in a book, you can use fine lines, small type, and subtle color differences. If that map will be placed on a website, design the map for a wide range of screen resolutions.
Resolution measures the smallest marks we are able to create within a display. It varies widely among the media on which we display maps. A desktop computer screen may show us 96 pixels per inch (ppi), while a high-definition (HD) TV with 1080p resolution can be about 33 ppi (because HD televisions have a fixed number of pixels, ppi varies with television size). A mobile device can show over 300 ppi. A laser print may squeeze 600 dots of toner in an inch to build the image. A litho plate on an offset press can reproduce 12,000 dpi from an image-set negative. Map features and type need to be larger to build them with pixels on a screen than to reproduce them in ink on a press. A map designed for screen display will look clumsy in a magazine, and a map designed for print may be illegible on screen. There are no bad media, just maps that are not designed appropriately for their media. Your map designs must change to accommodate each medium you are using.
When maps are produced in print-friendly formats, their resolution is set in dots per inch, or dpi. This refers to the number of ink dots that will be placed in a single inch. The file specifies this resolution to the printer, whether that is the laser printer in your office or an offset press at a professional printing house. The higher the dpi, the more ink dots the press or printer will place in a set area, making a crisper image. Obviously the dpi depends on the capability of the printing machine, but in general, it is best to set for-print dpi to at least 300.
A high-resolution file will typically come into graphics software with alarmingly large physical image dimensions, and the file may initially be many feet across when you were hoping to show it on letter-size paper. If you want to retain print resolutions, resize the file without resampling pixels (without lowering dpi).
For digital, on-screen display formats, however, resolution is set in ppi. The final appearance of a digital product depends heavily on the type of digital display device and its own supported resolution modes. Display resolution, image resolution, and image size are all linked together:
1. For an older computer display with a maximum supported resolution of 1024 x 768 pixels in a 14.2-by-10.6-inch viewing area, the resolution, also called pixel density, is 72 ppi both in the horizontal and vertical dimensions.
2. A larger monitor that displays 1920 x 1080 pixels with a 20.1-by-11.3-inch viewing area has a similar resolution of about 95 ppi.
3. A smaller tablet that displays 2048 x 1536 pixels with a 5.8-by-7.8-inch viewing area has a resolution of about 264 ppi. The higher pixel density on the tablet means that the pixels are physically smaller, and the device can show more of them in a single square unit of area. This results in images that appear crisper and more detailed.
Pixel density also affects visible image size. An image that is 360 x 360 pixels at 72 ppi shown on the 1024 x 768 display (1) would appear to be about 5 inches on a side as you look at the screen. That same image on the 1920 x 1080 display (2) would look about 3.5 inches per side because those 129,600 pixels fit into a smaller space, thus appearing smaller to the viewer. These changes in size change map scale—you have little control over the scale at which a reader will see your map on a computer screen or mobile display.
When exporting images for web or on-screen viewing, ppi does not mean much until you need to export images for multiple display resolutions. Changing the ppi for a digital image really means that you are changing the image’s pixel dimensions. For example, if you are using a monitor with a 96 ppi resolution, and you export your 360-by-360-pixel image at 96 ppi, the image’s final dimensions will be as you designed it: 360 x 360. That image will appear smaller on devices that have a higher display resolution or pixel density because those pixels can fit into a smaller visible area.
If you want your 360 x 360 image to appear at the same visible size, and equal sharpness on a phone with 384 ppi as it does on a computer with 96 ppi (a four times greater pixel density), you need to increase the pixel dimensions of your image by four times, to 1440 x 1440 pixels. If you simply enlarge the image with its existing low ppi settings and pixel dimensions, it will appear very pixelated (figure 4.1). An image set to a higher ppi than the display can support will simply appear at the maximum resolution of the display.
Although they are not the same thing (but roughly equivalent), dots per inch and pixels per inch are often used interchangeably, and different software packages will use different terms. ArcGIS uses the term dpi for all export formats, raster and vector.
Figure 4.1 A sample segment of a topographic map shown at 500 dpi (left) and 100 dpi (right). An inset enlargement of a 0 from the 600 contour label is shown at the upper right of each map segment. These enlargements allow each pixel to be counted: the zero is 45 pixels across 0.09 inches in the higher-resolution image and is 9 pixels across that same distance at the lower resolution. Source: Shetlerville quadrangle, Illinois-Kentucky, US Department of Agriculture (USDA) Forest Service DRG (digital raster graphic).
Figure 4.2 Land-use map of Tompkins County, New York. Map B shows a redesigned enlargement of the inset area (blue outline) from land-use map A.
Figure 4.3 Poor readability results when the enlarged inset is reduced in size and viewed at screen resolution.
Figure 4.4 The reduced map has been enlarged to demonstrate the pixelation that makes it unreadable.
Data sources: Tompkins County Assessment Office, Tompkins County GIS. Maps by E. Guidero, Department of Geography, The Pennsylvania State University (Penn State Geography).
Viewing distance affects map design just as resolution does. Features need to be enlarged to be visible from a distance. When they are enlarged, the resolution must be changed as well to avoid a jagged appearance. Letters 2 inches high that are seen at a distance of 14 feet are approximately the same size as 10-point type seen at a reading distance of 1 foot. A line 2 points wide is practically invisible from across the room, so line widths also need to be increased to retain visibility. (Points are a small unit of measurement used in graphic design; 1 inch contains 72 points.) Similarly, color differences need to be stark to make small features clear, whether they are small in measured dimensions or small because of the viewing distance.
The maps in figure 4.2 show land use in a portion of Tompkins County, New York (figure 4.2). A redesigned enlargement of the first map’s inset area (the blue rectangle) is shown in 4.2B. The enlarged map uses fine lines and small type that would be suitable for reproduction in print.
Simply reducing the enlarged inset to a smaller size demonstrates how the labels become unreadable with coarse resolution. There are not enough dots per inch to represent the small letter forms on this map at screen resolution, and lines have lost their detail and smoothness.
The odd-looking map in figure 4.4 shows the reduced version (figure 4.3) enlarged back to its original size with no redesign. You can see how poorly the type and lines are represented. You can also see how much information is lost at the coarse resolution.
Figure 4.5 is a redesigned version of the inset map made to display at this smaller size. Both lines and type have been enlarged, improving the legibility of a map this size (compare to the finer lines in figure 4.3). The larger version is redesigned again with large type and shown at a finer resolution in figure 4.6. This design would be awkward printed in a book to be read at close range, but it would work well for a poster intended to be viewed at a distance.
Both resolution of the media and viewing distance determine map design. These examples emphasize how type size and line width must change as a small part of a land-use map is resized and redesigned.
Thousands of colors can be produced in print and display environments. Printed pages do a good job of presenting color nuances. Maps created to be shown by projectors often require greater color contrast, especially when they include very light colors.
The flexibility of color selection varies widely with media. It is a good practice to test maps in the final media you intend them to be displayed. If the map needs to be readable in widely different media, produce different designs suited to these different contexts. If you want people to be able to make photocopies or prints of a map you are designing in color, test it out on a copier of moderate quality before you finish it. If you want a map to support a presentation, test it with a variety of brightness settings on a projector and look at it from far away with the room lights on. Make time to iteratively adjust colors and recheck prints of a map before including it in a paper report to a client. If you are going to spend $50,000 printing a book using professional offset lithographic printing, spend $1,000 on proofs early in the design process to check color sets used in the book. Looking at map colors on a color laser print is not an adequate check of how the offset printed colors will look. You do not want to leave readability of your important maps to chance. Testing how maps will look in their final form will help prevent many design disappointments.
Figure 4.5 The design has been improved for the small size. The inset is shown at the same scale as figure 4.3.
Figure 4.6 This design is suitable for viewing at a distance.
Data sources: Tompkins County Assessment Office, Tompkins County GIS. Maps by E. Guidero, Penn State Geography.
The maps shown in figures 4.7 through 4.9 were prepared from the same set of data. They show the change in number of crimes for local police beats in Redlands, California. Each map has a different purpose and thus has different constraints on color use. The first map has six color classes, ranging from dark to light to dark, through two hues (figure 4.7). This choice of colors emphasizes the highs and lows and provides details of change between the extremes. The two hues, blue and orange, represent decrease and increase respectively. The gray roads and white police beat outlines are base information that is readable but does not distract from the main message of change in crime. The readability of the white beat numbers relies on high-quality viewing or reproduction.
The second version of the crime map is designed for presentation on an LCD projector (figure 4.8). To anticipate differing qualities of projectors and different room lighting conditions, the map has been simplified to emphasize the highest increase and decrease in the area shown. This emphasis on extremes is supported with added text boxes that label the extremes. The beat outlines are also emphasized with a more intense color to ensure they are legible.
If the presentation graphics needed to be printed or photocopied in black and white, the dark orange and blue used in figure 4.8 would reproduce to grays that were too similar to distinguish. The same map redesigned once more uses only lightness to differentiate between increase and decrease in crime (figure 4.9). It is suitable for black-and-white photocopying or laser printing. Design constraints for black-and-white media that are unable to reliably produce shades of gray are particularly restrictive. Photocopying and faxing often restrict the mapmaker to black, white, and one or two middle grays for reproduction.
Figure 4.7 This crime map, which incorporates a detailed data classification scheme using many colors, is suited to high-quality display conditions.
Figure 4.8 The crime map now has a simpler classification with fewer colors and larger type, which would be appropriate for lower-quality display conditions of a projected presentation.
Figure 4.9 This grayscale version of the crime map is suitable for black-and-white reproduction.
Data source: Redlands, California, Police Department. Maps updated by P. Limpisathian, Penn State Geography.
Choosing appropriately among many export options lets people without GIS software view and manipulate your map files. It also allows a wide audience to see the maps you work hard to design.
A map is sometimes only one part of a larger presentation. To use a map in an online, digital, or print publication, you first must export it to a suitable graphic file format. When exporting a map, consider that it can be exported to one of two basic file format types—raster or vector—and can be used for two different display media—print or digital (on-screen).
A raster file uses a regular grid of small cells—pixels—to store color information across the map surface. It can be thought of as a picture of the original file. The size of cells in this grid determines the resolution: finer grids retain more detail but produce larger file sizes. Individual map elements, including text, are no longer grouped together as digital objects, but rather changed into collections of pixels. The file can only be altered by editing individual pixels.
A vector file maintains separable objects and renders their shape, size, and position in the file by connecting locations on the map. Even text characters are built from tiny curves connecting series of x,y locations with mathematical formulas. The degree to which map objects and text can be edited in a vector file depends on the file type chosen and its associated export options. Some vector formats may include raster elements as objects within a file.
There are trade-offs between quality, editability, and file size among all export file formats. As with many graphic decisions, testing the suitability of a choice before committing to it is an important step in producing a high-quality final product.
A simple map of Glacier National Park (figure 4.10) was exported from an ArcGIS MXD file to eleven other file formats. The resulting list of files and their sizes (which vary dramatically—from 322 KB to 11 MB) are shown in figure 4.11.
Figure 4.10 To compare export formats, a simple map of Glacier National Park (Glacier MXD.mxd) with just four line styles, five labels, and a terrain background was prepared.
Figure 4.11 Names, file types, and sizes for an example set of files exported from the Glacier National Park project.
Data sources: US Geological Survey (USGS), National Geospatial-Intelligence Agency/National Aeronautics and Space Administration (NGA/NASA), National Park Service (NPS), Canada Centre for Mapping and Earth Observation. Map by E. Guidero, Penn State Geography.
The three most common raster formats for exporting maps are JPEG (.jpg file extension), PNG (.png extension), and TIFF (.tif extension). JPEG is commonly used for web publishing because it uses a compression algorithm to store a slightly generalized version of the map in a smaller file, which makes the image load and display faster. PNG is also commonly used for web publishing and uses a different “lossless” algorithm to avoid compression artifacts (shown later in the chapter). TIFF produces pixel-by-pixel renditions of the map and has largely replaced BMP (bitmap) format. TIFF is recommended if you need to export an image for print because the format is widely recognized by multiple operating systems and software packages. It contains the most pixel information possible, especially when uncompressed. PNG and TIFF support both kinds of transparency (index and alpha), whereas JPEG does not.
Resolution options vary among formats. To export to most raster formats, height and width dimensions in pixels, or a set resolution in either ppi or dpi, are specified. In ArcGIS, the default export resolution is 96 ppi (suited to desktop computer screens). However, 96 dpi results in a coarse printed image because printers and presses can render a much higher dpi than a computer display. When the map of Glacier National Park was exported to TIFF, the default resolution choice (96 dpi, 432 × 504 pixels) produced a file with low resolution but a relatively small file size (639 KB). An enlarged section of the low-resolution TIFF image is shown in figure 4.12.
Exporting the map again to TIFF but with a 400 dpi resolution (1800 × 2100 pixels) produces a higher-quality image (figure 4.13). The type edges are no longer jagged, and the lines are much smoother. The pixels making up this image are about 1/16 the area of pixels in the coarse-resolution version when they are examined at the same map scale, and there are many more of them. Smaller features can be recorded at higher resolutions.
Higher image quality comes at a price. The high-resolution setting produces a much larger file—seventeen times larger than the coarse version. To put this size difference in context, if you had a 1 GB storage limit, you could store more than 1,600 coarse-resolution files, but only 94 high-resolution files.
Other choices for file export, beyond numbers of pixels, will also affect file sizes. The color-depth setting you choose affects file size, with 24-bit files producing the largest exported files. Higher depth numbers provide greater numbers of distinct colors in an image because more digits are allotted for storing color information for each pixel in the file. Some formats also allow you to choose compression type or level. For ease of use and highest-quality TIFF images, use the LZW (Lempel-Ziv-Welch) compression method. LZW is a lossless compression format, meaning that an LZW-compressed file still contains all the information of the original, but is small enough that it can be readily shared. The TIFF example shown in figure 4.13 was exported at 400 dpi and left uncompressed. It is a high-quality image suitable for publication. Again, this quality came at the price of a large file size (over 11 MB). With LZW compression, the file size reduces to 2.7 MB. Even with compression, TIFF creates images that are generally too large to be displayed on the web.
Figure 4.12 A portion of the exported Glacier National Park TIFF at 96 dpi. You can clearly see the pixelation that the low resolution produces.
Figure 4.13 A portion of the Glacier National Park map exported as a high-resolution TIFF (400 dpi).
Data sources: USGS, NGA/NASA, NPS, Canada Centre for Mapping and Earth Observation. Maps by E. Guidero, Penn State Geography.
JPEG formats use a sophisticated compression algorithm to make high-quality raster files smaller. JPEG is a lossy compression algorithm, which means that data is lost when a map is exported as a JPEG. If you lower the quality of the JPEG compression, be sure that you will not want to enlarge the resulting image in the future. If you do enlarge the image, you will see compression artifacts, or areas that look like a messy haze around higher-contrast elements (figure 4.14).
When exporting to JPEG from ArcGIS, you can set the resolution (in dpi) and the level of compression, called quality (ranging from low to max). Maximum quality (level 100) is the minimum level of compression, resulting in a larger file size but a much higher-quality image. Lowering the compression quality increases the level of compression, which decreases the file size but also the quality of the image.
You can see the difference this compression makes by comparing the maximum-quality and medium-quality exports. The map shown in figure 4.15 was exported to a JPEG at 400 dpi and maximum quality (Glacier max-qual JPG, 2.2 MB). The result is a high-quality image with a reasonable file size. This size is too large for web display, but the quality is good enough that it could be used for some printed contexts. The map shown in figure 4.14 was saved as a JPEG at 400 dpi with medium quality (Glacier med-qual JPG, 322 KB). The savings in file size is very good but notice the speckled artifacts around lines and type. The reduced file size savings comes at the cost of a poorer-quality image. The file would be fine for many web applications or Microsoft PowerPoint presentations, but the quality is not good enough for print publication.
You can also compare the max-quality file (figure 4.15), which still contains some compression, to the high-resolution TIFF image exported at the same resolution (figure 4.13). The TIFF is over twenty times larger.
Figure 4.14. A portion of the Glacier map exported as a medium-quality JPEG file to show hazy compression artifacts enlarged (400 ppi).
Figure 4.15 A portion of the Glacier map exported as a high-quality JPEG file (400 ppi).
Data sources: USGS, NGA/NASA, NPS, Canada Centre for Mapping and Earth Observation. Maps by E. Guidero, Penn State Geography.
The PNG format does compress images, but uses a lossless algorithm, which eliminates compression artifacts. Figure 4.16 shows a PNG image. PNG file sizes are larger than JPEG (but smaller than TIFF).
All raster file formats share the common characteristic that they are made of only pixels. In raster files, text, lines, and colors are difficult to edit. For example, editing a label would require using graphics software to erase the existing pixels that form the characters and then overlaying new text. Lines and areas do not continue beneath the text and would need to be repaired as well. Changing the font or style of many map labels requires going back to the GIS software and re-exporting the map because there are no text objects on the raster map to select and change. To change a color in the map, you would need to first select every pixel in the area to be changed, working around text and lines that overlay the area. This can be done, but it is much harder than selecting polygons in a vector image and making the color change once.
These raster export formats should be used only for maps that you want to show or print “as-is.” This inflexibility can be an advantage when you do not want to pass on a version of your work that can be easily edited or adapted for other purposes.
The three most common vector formats used to export maps from ArcGIS are AI (.ai extension), PDF (.pdf extension), and, to a lesser extent, SVG (.svg extension). PDF has largely replaced Encapsulated PostScript format (EPS, .eps extension), although EPS export is still available and can be used. SVG is an XML-based open standard, and thus, like PDF, can be used to transfer vectors between different programs. Vector files are often, but not always, much smaller than the raster files discussed in the previous section. Comparing exported vector files of the Glacier National Park map, the AI file is 9.2 MB, the PDF is less than 1 MB, and the SVG file is 2.4 MB. Recall that raster files of comparable quality were 11 MB (figure 4.11). An elaborate file with many small features, complicated line work such as detailed contour lines, and numerous labels could easily produce a vector export much larger than a corresponding raster export.
Note that when exporting to these three vector formats, any layers underneath a transparent layer may automatically be rasterized. If you wish to keep certain vector elements as vectors in the export, reorder your map layers before export.
Figure 4.16 A portion of the Glacier map exported as a PNG file (400 ppi).
Figure 4.17 A portion of the Glacier map exported as an AI file. Notice the word “Glacier” does not maintain the correct halo positioning.
Data sources: USGS, NGA/NASA, NPS, Canada Centre for Mapping and Earth Observation. Maps by E. Guidero, Penn State Geography.
Exporting to the AI format results in complete text strings and high-quality Bézier curves for lines (splines). For example, “Glacier” (as part of the label Glacier National Park) can be selected as a single object, then edited or restyled (figure 4.17), and the path it follows can be easily manipulated. This is important because you want the process of editing text labels to be easy and not require tedious repairs or manual replacement of individual letters or words. Map elements exported as vector objects can be altered in shape, texture, and color in Adobe Illustrator or in other illustration software that can import AI files.
Because ArcGIS often exports to an older version of the AI file type, there are a few downsides to exporting to AI. Many layers, groups, and clipping paths are generated in the AI export, and often labels are grouped in odd ways. Selecting “Glacier” also selects the phrase “Waterton Lakes National Park.” You have to expand the layer structure to isolate and edit “Glacier” or use other tools to access individual lines of text. Curved labels are also often positioned incorrectly, and halos are not registered to their labels. Notice in figure 4.17 the letters for “Glacier” are no longer registered with their halos. Halos are used to make text more legible over the top of complex line work or other backgrounds. Each halo is exported as an individual vector object, and exporting many halos can create an unmanageably large file. It may be easier to apply halos, also called strokes, in the post-GIS editing phase.
Figure 4.18 A portion of the Glacier map exported as a PDF file. Individual letters are selected (faint blue lines) to show that the PDF export does not group letters as whole words.
PDF is a common, supported format that is editable in graphics programs. Opening the PDF in Illustrator still allows text to remain complete and editable (figure 4.18). Unfortunately, type strings are broken up into individual letter segments, which means you should not rely on the post-GIS editing to fix problems with labels and type because it will be time-consuming to make changes at this stage. PDF includes the same layers, groups, and clipping paths as the AI export does.
ArcGIS provides many options for creating a PDF file from the map document. Compression options for vector and raster components can each be set individually. ArcGIS also offers the option to include feature attribute information from each layer’s attribute table. While this makes for a much larger file size, it may be useful to do if the end user needs access to feature information but does not have GIS software.
The SVG export, like PDF, breaks text up into individual letter objects, so any text styling needs to be applied to each individual object. Halo registration and letter spacing are also severely compromised. For example, the spacing between letters in words is far larger than the spacing in the initial file (figure 4.19).
Figure 4.19 A portion of the Glacier map exported as an SVG file.
Data sources: USGS, NGA/NASA, NPS, Canada Centre for Mapping and Earth Observation. Maps by E. Guidero, Penn State Geography.
Despite the problem with halos and excessive layer creation, the AI format will be the most trouble-free for publication. You should be aware that some custom type effects and special characters do not export well; therefore, always test your choices before relying on them for a design that will need to move beyond GIS software.
Transparency is a useful image property but can cause problems in your output if you do not pay attention. Transparency becomes especially important when creating tiled maps or bivariate maps and legends. Transparency can also be used to good effect in printed maps to give the impression of overlays. If you choose to have transparency in any layer, you need to select an output format that can support it. TIFF, GIF, and PNG support transparency, but JPEG does not. If you export a simple map that contains a transparent layer to JPEG, your image will be flattened. The resulting map will appear as if there were a transparent layer, but in reality, the entire image was simply converted to the equivalent RGB values for each pixel.
If you are working with tiled maps, you may want a layer that has transparency or is semitransparent to function as an overlay through which you can see basic reference information in the basemap. For example, you may want to see highways and city names through a map of precipitation ranges. If you need transparency, you will need to cache your map in the GIF (graphic interface format) or PNG format—there are different subformats of PNG that cover different bit depths and transparency types. PNG-8 and GIF can display 256 colors (8-bit color), and support index transparency, which means that map areas are either fully transparent or fully opaque (figure 4.20). Here, in GIF, the land is completely transparent and the lake is fully opaque. An 8-bit image results in a fairly small file size and would be useful for data or overlays that themselves do not need be transparent, but have transparency wherever data is not present. For example, a layer showing urban footprints may have no content outside city extents.
PNG-24 (24-bit color) can display 16 million colors, called true color, but like PNG-8 supports index transparency only. PNG-32, which supports 24-bit color plus alpha transparency, allows for varying percentages of transparency, which is useful for data overlays through which the basemap content needs to be visible (figure 4.21). Here, the lake edge is 75 percent transparent. This halo could not be included in the GIF due to lack of support for partial transparency. PNG-32 takes up the most space, closely followed by PNG-24, because it includes so much color and transparency data. You may want to avoid using either the 24- or 32-bit formats unless you absolutely need more than 256 colors or if you have complex transparencies or transparent gradients.
Figure 4.20 PNG-8 and GIF, 8-bit color formats, support index transparency.
Figure 4.21 A 32-bit format will support alpha transparency, which allows for any grade of transparency between 0 percent (opaque) and 100 percent (fully transparent).
Figures 4.20 and 4.21 by E. Guidero and P. Limpisathian, Penn State Geography.
There are three export formats well suited for displaying static map images on the web: JPEG or PNG for simple, uneditable images, and PDF for higher-quality maps that contain both vector and raster elements, which can be turned on or off at the viewer’s discretion.
PDF, as was discussed previously, is a high-quality vector format common for web display and distribution and can include both vector and raster layers. PDF files can be viewed with Acrobat Reader, which is available free from Adobe. PDF works well for displaying a high-quality vector image that can be panned and zoomed. The file with a segment shown in figure 4.18 is 878 KB.
Although PDF is an editable format, because of the problems associated with letter-by-letter editing (discussed previously), you should use PDF for maps with a completely finished design. Furthermore, if you would like your map content to be searchable (that is, if you want the text to be fully indexed and searchable by a web browser), during the export process in ArcGIS, you can enable the export of georeferenced information. For the map title and other metadata to also be searchable, remember to add the title and any additional information, including metadata, to the properties of the map document itself prior to exporting.
Caching is the process of creating small square rasterized pieces of a map, called tiles, which seamlessly fit together for display in a zoomable and pannable map on the web (sometimes called a slippy map). A tiled map will have several discrete scales, and each scale contains a different set of tiles. As the user zooms in and out, the set of tiles for that particular scale is loaded. Once each set of tiles is loaded into memory, cached maps display very quickly. Your choices when creating maps for caching affect the total cache size, upload time, and time to generate the tiles.
Each additional scale, or set of tiles, will increase the size of a cache. As you prepare to cache a map, consider whether your map layers need to be visible at all scales. Consider also at what particular scales you want to generate map tiles. The traditional tiling scales are those used by Google Maps, Bing Maps, and ArcGIS Online (and those, in turn, have set the standard for most other tiled maps). These scales span approximately twenty levels, sometimes called zoom levels, starting at 1:591,657,551, level 0 (see table 4.1). The scale factor increases by two for each successively larger scale or level (that is, the denominator of the representative fraction is halved). The largest scale commonly in use is 1:1,128 (level 19), although other web maps can include scales larger than that. Recalling chapter 1, these are scales at the equator with a web Mercator projection, not the scale at other latitudes of views seen on screen.
Table 4.1: Zoom (caching) levels and associated scale at the equator
ZOOM LEVEL |
SCALE |
0 |
1:591,657,551 |
1 |
1:295,828,775 |
2 |
1:147,914,388 |
3 |
1:73,957,194 |
4 |
1:36,978,597 |
5 |
1:18,489,298 |
6 |
1:9,244,649 |
7 |
1:4,622,325 |
8 |
1:2,311,162 |
9 |
1:1,155,581 |
10 |
1:577,791 |
11 |
1:288,895 |
12 |
1:144,448 |
13 |
1:72,224 |
14 |
1:36,112 |
15 |
1:18,055 |
16 |
1:9,028 |
17 |
1:4,514 |
18 |
1:2,257 |
19 |
1:1,128 |
Think about the scales you really need—if you are using municipal-level data, you likely will not need to show it at anything smaller than 1:36,112 (level 14). Likewise, if you are intending to display global-level data, you probably will not need to create tiles at scales any larger than 1:18,489,298 (level 5). Using a standard web Mercator projection on a 24-inch monitor with 1920 x 1080 resolution, the smallest scale—level 0—is enough to show the globe multiple times over. You also have the option to create tiles at any scales you wish. If you choose nontraditional scales, it may be hard to combine your maps with other web content and services.
Transparency is a key issue in choosing a file type for caching (see previous section). Do you need tiled layers to be transparent so other content can be seen through them? Increasing quality and transparency capabilities of image data creates larger caches. Larger caches can create issues with processing time, storage, and display responsiveness.
If you are creating your map or graphic to be displayed on a US government website, be aware of requirements to comply with Section 508 standards. Section 508 is a set of laws governing the accessibility of electronic information by people with disabilities. Many large institutions, such as universities, also adopt these standards for their web content and educational materials in online course resources. The full set of standards can be found at http://www.section508.gov/content/learn/standards. This set of laws includes standards for the presentation of graphics and images on the web. With respect to maps in particular, maps with animation and color-coding must also be made available in nonanimated and non-color-coded versions (that is, black-and-white or textual descriptions). Chapter 8 provides information on designing for color-blind readers.
Maps, charts, and globes are all protected under copyright law. As a cartographer, you have copyright protection for the maps you produce, just as others hold the rights to their own original work. Violation of copyright protection can result in serious legal charges. Understanding copyright and licensing will allow you to proceed with confidence when sharing or using maps and data.
Copyright is a set of laws that protect the exclusive rights of a creator to reproduce or publish original work. Anyone who produces something original, such as a map, has the right to say how that work should be published or reproduced. They can also exclude others from using the original work. It is the idea of original, creative work that is critical to understanding what is under copyright protection.
In cartography, the line can seem unclear for what is considered original and creative work. You can think of cartography as a blend of geographic facts and the creative expression of these facts through design. Copyright protection applies only to the expression of these facts, not the facts themselves. No one can own exclusive rights to the shape of a river bank or location of a town, but the way these geographies are generalized and symbolized on a map does fall into the realm of creative expression and is under copyright protection.
A map cannot be under copyright of the creator until that mapmaker puts an “appreciable” amount of creative work into the product. Someone making a map derived from various data sources should not claim the product as their own until they have put enough effort into the design of the map, including colors, symbols, fonts, selection of data, and generalization. For example, if you trace the road network from a published map for direct use in your own work, you could be infringing on the rights of the original producer, who assumedly put an appreciable amount of work into the generalization and selection of that road data.
Alternatively, you may be putting an appreciable amount of effort into designing a map as an employee of a company or person. Your map, however, is likely a “work for hire,” and thus the publisher owns the copyright, rather than you as the mapmaker (see “Work for Hire,” later in this chapter). As an aside to professors, university intellectual property rules often place copyright with the student for course projects. Treat students’ projects as copyrighted—gain permission and credit authorship when you seek to use their great maps in publications.
Work for Hire
As an example, core portions of the memorandum of understanding (MOU) for the student cartographers who assisted revision of this book read as follows:
Dr. Cynthia A. Brewer, through her appointment by the Department of Geography of The Pennsylvania State University, is employing you, independent of any coursework and not for academic credit, to assist in the revision to the first edition of the book “Designing Better Maps” with the expectation that the second edition will be published by Esri Press. . . .
Your work on this revision will be completed as a work for hire, under 35 U.S.C § 101 and as further defined by the US Copyright Office Circular 9 (www.copyright.gov/circs/circ09.pdf), which states that “[i]f a work is made for hire, an employer is considered the author even if an employee actually created the work. The employer can be a firm, an organization, or an individual.” Consequently, the copyright for any and all work created as a result of this employment shall be owned solely by Esri.
While creative products can be under copyright, the specific techniques used to create these works do not have copyright protection. A classification technique or smoothing algorithm is not under copyright, but the results of the technique, when applied creatively, may be copyrighted.
Original work is assumed to be the copyright of the creator with all rights to reproduce or publish reserved to the creator of the work. There does not need to be a label saying this work is under copyright for copyright law to apply. To share the rights to reproduce or publish, the creator can enter into a licensing agreement with individuals, organizations, or the public.
Licensing allows the creator to share some or all rights to the work with a licensee. In a license agreement, the creator maintains copyright of the original work, but the licensee is now allowed to use the work in some way. For example, a firm may license their maps to another group to allow for their reproduction, often in exchange for a fee. Licensing agreements will typically include stipulations on permitted use, limitations, collaboration allowances, restrictions on users, and reserved rights. You or your company might use a signed permissions form to offer nonexclusive use of copyrighted material in your maps. To reduce the administrative work of obtaining copyright permission, most maps in this book are original works. For maps by others, I gained permission for use in the book with a corporate copyright permissions form. This is an important step in publishing a commercial product with example maps in it.
Cartographers often have additional graphic elements as part of a map product, such as a photograph of a distinctive habitat or welcoming downtown. Do not neglect gaining copyright permission for these—photos are also creative works protected by copyright. Photographers sometimes use a common license if they want to share their work (described in the next section). You should not assume that a photo freely available on the web (of which there are millions) is also available for use in your commercial map product.
Cartography firms sometimes add copyright hooks or traps to their map documents. These may be small streets or extra locations that do not exist in the real world and that will not lead the map reader astray. Hooks allow the company to unambiguously identify a derivative product if they need to press a case of copyright infringement.
Be careful that you do not misinterpret the concept of “fair use” as a way to avoid copyright constraints. Fair use allows reproduction of a small part of a work for criticism, teaching, research, and other purposes (http://www.copyright.gov/fls/fl102.html). Fair use court decisions are affected by whether the use has a commercial purpose, the proportion of the whole used, and how the use will affect the market for the original work. For copyrighted text, quoting short passages that are very small proportions of the whole work is considered fair. The ways fair use may apply to the graphic detail of a map are difficult to interpret.
Typically more useful to a cartographer are public licensing agreements. As the name implies, these are licenses that transfer some or all rights to the general public. Organizations such as Creative Commons (CC) have developed licenses that are free to use—as long as you include the appropriate credit line—and help both creators and users of material manage copyright and licensing for the general public. The Creative Commons has four common license features that are combined into six unique licenses, which can be applied to creative work. All licenses use Attribution, meaning users of the creative work must give the creator credit. A license may use the NonCommercial feature, which lets users share, distribute, or modify the creative work as long as it is not for a commercial purpose. Licenses may also include provisions for either ShareAlike, meaning derivatives and distributions of the original must be shared under the same license as the original, or NoDerivs, which prohibits modification of the licensed work. For example, Attribution-NonCommercial-Share Alike means you can share and change the work as long as you attribute your source, you are not selling the product, and you require derivatives to be under the same license as the original. Creative Commons provides detailed descriptions of each license at the Creative Commons website, http://creativecommons.org/licenses. Figure 4.22 shows the symbols used to indicate each license feature.
Figure 4.22 Creative Commons license types. Source: Creative Commons, https://creativecommons.org/licenses/by/4.0/.
One form of sharing work with the public is open data. Open data is based on the idea that data should be freely used and shared by anyone without restrictions of copyright. Open data is intended to be accessible and available with no cost beside what is reasonable for reproduction. Users of open data have the right to reuse and redistribute the data, and no one is prevented from using the data. Ideally, open geodata is data available for free download from the Internet and is provided in a nonproprietary format such as comma-separated values (CSV) for tables, TIFF for rasters, or shapefiles and GeoJSON for vector data. License agreements even for open data or “free” data often hinge upon the user including a credit line. It is always best practice to acknowledge data sources for your maps.
Creative works that are no longer under copyright protection have been released into the public domain. These works are free for anyone to use with no limitations and no legally required attribution of the source (although attribution can be informative, and giving credit is a courtesy you should show whenever possible). All work and data produced by the US federal government and all work and data published on the Natural Earth website (http://www.naturalearthdata.com) are in the public domain, which is useful to cartographers. The Creative Commons organization encourages use of a “no rights reserved” mark to clarify that a product is in the public domain (figure 4.23).
Because US federal government products are in the public domain, some mapmakers err in thinking all government data is public domain. The national map products of other countries may not be public domain. The maps and data of local governments in the United States are also not public domain by default, even if they are made freely available on the web. A town, city, county, state, or other local government agency may want their community to access spatial data and maps for decision making, but that does not mean they want it to be the content of another’s commercial map venture. Permission, permission, permission.
The widespread availability of maps and data can sometimes make it dangerously easy to forget about ownership and violate copyright law. As a cartographer or GIS professional, being cautious and aware of copyright when using outside maps or data sources is important to avoid legal infringement on the rights of others. When publishing data or maps to the web, you automatically have copyright of creative original work, but the best practice is to make clear under what license you intend the work to be used or shared, if at all. Provide your contact information with your maps and data for those who seek permission to use them.
Figure 4.23 The CC0 mark recommended by Creative Commons for public domain products with no rights reserved. Source: Creative Commons, http://creativecommons.org/licenses/by/4.0/.
