Cartography

Metadata

• Author: Kenneth Field
• Full Title: Cartography
• Category: books

Highlights

• Because the world around us is a complex one, it would be virtually impossible to simply place a small version of it on a map. There would not be the space to adequately represent all features that exist in the mapped area even in a reduced form. Consequently, all maps are abstractions of reality and are used to display a selection of objects and attributes. All maps are inherently a reduction of reality and so the amount of information you can put on a map will be a reduced form of that reality. This means a map will omit information to a greater or lesser extent depending on scale and purpose.
  But more than reduction (through selective omission), the features that are mapped are subject to a range of additional processes, such as classification and simplification, that make it easier to understand the true spatial patterns and relationships that exist in reality.
  The way in which we represent features and their attributes is through the design and placement of graphical signs.
  These signs do not necessarily take on the appearance of the object in reality but are used to represent the object. The signs should have meaning to enable the map reader to interpret them accurately and appropriately relate them to the realworld object. We refer to the process of encoding meaning into the map as symbolisation. At its simplest level, part of the job of the mapmaker is to design and place symbols that reflect either location or some characteristic of the data. There is considerable scope in the design of symbols and every mark on a map can be considered a symbol in one form or another, from those that represent points, lines, and areas to the typographic components and the marginalia or contextual information.
  Clarity and purpose follow from a careful consideration of abstraction and signage. Beyond the science of choosing and representing the information, the sophistication of the intended audience, scale, and conditions of use must also be considered so you end up with a clear and concise map that your map reader can easily translate into meaning.

New highlights added February 24, 2024 at 2:39 PM

• Surfaces represent the geography of values that exist across an entire region. Although isarithmic maps allow you to create contoured representations of a surface, they classify the data into bands. Instead, a continuously varying surface might be used. Continuous surfaces are interpolated from data that is collected at sample points either by mapping the density of points or the values they represent. Continuous surfaces are good for showing data that is known to vary gradually from place to place rather than having abrupt boundaries.
  Sampling data is usually used for continuous phenomena or for data collected at points. Spatial samples can be random where sampling points are unevenly distributed across space.
  Alternatively they might be systematic, or regularly spaced, or stratified, where more sample points are used in some areas than others on the basis of some prior knowledge of the feature's distribution.
  Creating an interpolated raster-based continuous surface requires statistical analysis of the input data. A variety of methods can be used. Exact interpolators include the actual data values as values in the final map. Smoothing interpolators create a more generalised surface and do not necessarily include actual data point values in the final map. Common methods include Inverse Distance Weighted and Kriging, of which each has multiple variants depending on input parameters to the algorithms. There are many alternatives diffusion methods. (Page 74)
• Contours are lines that connect points of equal value, such as elevation, bathymetric depth, temperature, precipitation, pollution, or atmospheric pressure. In this sense they can be used to map more than just elevation and map various thematic datasets as well. The distribution and separation of the lines shows how values change across a surface. Where there is little change in a value, lines are spaced farther apart.
  Where values rise or fall rapidly, the lines are closer together.
  Contours are usually constructed from spot height information or a digital elevation model (DEM). The base contour is the value from which to begin generating contours.
  Contours are generated above and below this value as needed to cover the entire value range. The contour interval specifies the distance between contour lines. The choice of contour interval and smoothness of the line's curvature is a matter of choice led by the amount of other content on the map and the map scale. You'll likely want to generate a more detailed set of contours for larger scale products.
  Contours can be symbolised in many ways. Typically these include using a single colour (commonly orange/brown) or a colour ramp that shows lower contours as darker and higher contours as lighter. Value can modify hue when symbolising bathymetric contours. It is also common practice to change the thickness of the lines you might want to use as interval (or index) contours to emphasise them. Interval contours usually represent major values in elevation-for example, every 1,000 m where intermediate contour lines might represent each 200 m in between.
  Sometimes, c modified rep For instances like features among other small rightto the down to focus att distinguish depressio Not every contour line needs labelling. After all, these are already abstract representations placed on the map so adding too many labels might clutter the final image. It is usual to label interval contours. Positioning is usually in a stair-step placement pattern around other map detail to assist interpretation of slope direction. Alternatively they can be right reading. (Page 78)
• Copyright laws exist to protect. It's the legal right that grants you or the producer of the work exclusive rights to its use and distribution, often with some limitations. Different copyright laws exist in different countries and as much as they protect you from having people breach the use of your work, you must also ensure that you are in compliance with the copyright on what goes into your map. Limitations might include fair use or the fact that it covers the expression of an idea and not necessarily the underlying ideas themselves. Put simply, you cannot copyright a 'fact' but you can copyright an expression of that fact which makes it easier to distinguish the more original content you have created.
  Copyright is one form of establishing intellectual property for creative works. Maps can fall into this realm, which offers protection to the original authors of such works. For instance, in the United States copyright includes placing restrictions on the rights of others to reproduce your work, to prepare derivative works based on your work, or to redistribute, sell, or transfer ownership of your work. Copyright can be officially registered, or you can simply add an appropriate statement on the map asserting your copyright. Copyrights are territorial, and although international copyright agreements exist, it is worth checking the copyright laws of different countries.
  Maps and data you use are also likely to have some copyright associated with them even if you acquire them for free. It is important to establish what copyright and licensing restrictions exist on the raw map materials you use.
  Many datasets may be free from copyright but it shouldn't be presumed. Even when you are free to use data there are often statements of compliance that you must include. Always request permission from the copyright holder if you intend to use a part of their work in your own work. Permission can usually be granted in writing or through a contract that specifies quantity, royalty, or licensing fees and any specific terms. Failure to gain permission can result in penalties including payment of profits or damages, court costs, and lawyer's fees. The biggest cost is in terms of credibility. (Page 80)
• Make sure your map is accurate' would seem like sound advice but what, exactly, does accuracy mean? Accuracy refers to the quality of being correct or exact, a description of how close the map is to reality. Accuracy includes the quality of the data used to make the map. Are locations or magnitudes correct or, perhaps, are there errors in the dataset that means the resulting map propagates error, leading to uncertainty?
  Accuracy shouldn't be confused with precision, which refers to the exactitude of the mapped features. For instance, the level of measurement of a location or empirical values play a significant role in determining how precise a value is. Precise location may measure an x,y paired coordinate to a fraction of a unit and, so, locating a position to a centimetre is more precise than a metre. Measuring magnitudes to several decimal places gives a greater level of precision to the same values rounded to an integer.
  Data can be accurate and imprecise or, conversely, precise but inaccurate. They are two separate measurements, both of which can have inherent errors and which build a picture of the overall quality of the data. You may not have much of a say about the precision of the data you use for mapping if you've not been involved in its specification or collection. You may decide to use a less precise form of the data for your map by, perhaps, rounding numbers to create a more generalised form.
  It's worth remembering that you generally cannot make a map that purports a greater degree of accuracy or precision than the data supports. There is usually a minimum scale at which the data holds up to scrutiny. Purporting to show data at a larger scale than its quality supports should be avoided. For instance, thematic data collected at one level of spatial aggregation cannot be imputed to smaller level boundaries. This is often referred to as the 'ecological fallacy-the interpretation of statistical data in which inferences about individuals or smaller areas are erroneously deduced from the group to which the individuals or smaller areas belong. (Page 96)
• Geographical phenomena can be arranged along a discrete-continuous arrangement. The terms 'discrete and continuous are often used to describe different types of data along a numeric line but they can also be used in cartography in a spatial context. Discrete phenomena occur at distinct locations and are differentiated from other discrete phenomena by the intervening space. For instance, at any moment in time, each individual in a town occupies a discrete location that cannot similarly be occupied by another person.
  On the other hand, continuous phenomena occur throughout a geographical region. Elevation is an example of continuous phenomena because every latitude and longitude position has an associated value of elevation-there is no empty geographical space between values of elevation for which there is no value of elevation.
  Both discrete and continuous phenomena can also be arranged as abrupt or smooth. Phenomena that change suddenly are abrupt whereas phenomena that change gradually over geographical space are smooth. For instance, voting patterns for electoral constituencies are likely to change abruptly between neighbouring areas. In fact, this is an example of abrupt continuous phenomena because electoral constituencies exhaust space and there is a value associated with each one, but an abrupt change occurs at boundaries between constituencies. An example of smooth continuous phenomena would be the measure of total precipitation over the course of a year. Abrupt changes in precipitation are not likely across geographical space.
  Different map types are more appropriate than others depending on the arrangement of data. Selecting a map type that supports the data type and arrangement will help convey the information appropriately although it is possible to convert between data arrangements, for instance by
  aggregating from continuous to discrete or binning smooth to abrupt. (Page 98)
• Descriptive maps show where phenomena are located, organizational structure, subdivision, or routing. They are explanatory and non-quantitative but generally go further than the content of a topographic reference map. The detail on a descriptive map is usually very specific. For example, power lines or sewers, sales territories, or the location of stores. The specific theme is generally mixed with other contextual topographic content so the map becomes neither a pure topographic nor thematic map. The topographic detail included is normally only the bare minimum needed to support the contextualisation of the theme being described.
  Information is encoded using typical map symbology though industry-specific styling and symbology are often used. The inclusion of quantitative data is generally tangential to the main theme and might be included using marginal elements such as tables or graphs. Quantitative data is not the primary focus of the map. Scale can vary depending on the purpose from small, large-scale maps showing a store location to large, small-scale maps that illustrate the location of all franchises in a company on a global scale.
  Another way of thinking of a descriptive map is that it is a hybrid. It takes some of the skeletal aspects of a topographic map but incorporates an additional theme. The theme is unlikely to be understandable as a purely thematic map because it's qualitative so it needs some topographic detail. The map is focussed rather than general purpose. (Page 114)
• At one time or another, most people will have read a map yet they come to that task with many different and varying experiences. Unless you are making a map for an extremely niche readership, you're probably going to face the difficulty of trying to author your map to suit a number of different use case scenarios and different map readers. There are ways you can organize the map to feed off people's awareness of other media.
  People generally feel comfortable with the familiar. In the same way that they are likely to have read a map before, they will also be familiar with the structure of their favoured newspaper, websites they use, and the multitude of different forms we have to complete for banking, health care, taxes, and so on. They all embody very different characters. Some are impersonal and indifferent; others are empathetic. They are designed to elicit a particular response. This characteristic der be mimicked in the design of maps.
  Hipping for neutrality and indifference is actually quite hard. Perhaps you're mapping simple but pure numerical information, and the map's function is simply to share that pattern or relationship without really caring whether the user needs to be concerned, act upon the results, or deal with the consequences. Such a map can be faceless in the sense that it doesn't need to evoke a response. Fonts and colours can be muted. Layouts can be relatively flat and systematic. You want to give the reader enough to understand the map but without it evoking personal feelings for the subject matter.
  Conversely, you might want to evoke a response. This might be to lead the reader to action or comment; or support them in doing something as a result of the map. In this sort of map, you might emphasise certain characteristics. You might use colours and fonts in a more dramatic way to evoke certain feelings. Setting out to evoke something particular in a readership has every chance of missing its mark. Fully appreciating the user needs and the persuasive context of your map is vital to success (Page 116)
• 120 many Maps don't exist without data, and the web has supported y dramatic changes in the availability of data. Beyond traditional data collected and stored by mapping agencies, much new and interesting data has become emancipated from silos and been made available for people to download and map. The public has also been instrumental in building new datasets and sharing them directly with each other, or augmenting traditional data. For example, our ability to locate ourselves and record information has led to some paradigm shifting datasets such as OpenStreetMap which rivals and often exceeds commercial data as a complete and contemporaneous dataset which supports community mapping as well as many commercial ventures.
  The growth of citizen science and our ability to act as a sensor-network to capture, record, and share volunteered geographic information has supported the explosion of online thematic mapping. Data is now published rapidly, often with no extra effort (it is passively just shared), and is either already georeferenced or can be geocoded easily and consumed on demand. Cartography used to be based on meticulous surveys that took days, weeks, and months before being wrangled onto a map. It was often out-of-date before it had been published. Data is now streamed live from sensor networks, consumed inside maps automatically, and published in real time to give us a truly on-demand live picture of the world around us.
  The change in data capture, handling, and availability has had profound effects on the cartographic process. Where once, data was only accessible by a few and cartographers were key to the delivery of maps based on it, now many can access the same data that professional cartographers use. This has increased both experimentation and the sheer volume of maps as well as the number of people making maps. Data will continue to be the bedrock of cartography.
  Maps will always be required as the principal mechanism to communicate the spatial essence of data where geography is a fundamental component (Page 120)
• A good map embodies a functional balance between the information represented, the skill of the cartographer in revealing that information, and the context in which the information is viewed. This includes the background, education, and experience of the map reader. Getting these components to balance is where the profession of cartography comes in. The cartographer sits at the fulcrum of the balance.
  Constraints can lead to design and representation issues that must be handled properly to avoid creating a dysfunctional map; one in which some basic rules might be ignored or where the design fails to fit what we know about how maps are read and used. There are any number of reasons why the best approach may be not to make the map at all.
  Data is often a principal cause for making a dysfunctional map. Data may be incomplete, skewed, or missing. It may lack meaningful detail or just be too oversimplified to support the idea behind the map you want to make. It may be that the data isn't able to reveal the meaning you'd like or that it is too detailed to be transformed into something understandable.
  Not understanding data can also lead to unintended.
  consequences. The cartographer has to be aware of when their own limitations impede the design. Being unable to distil data into meaning isn't a crime-but making a map of something you simply don't understand can potentially propagate misinformation.
  If the objective of making the map is to simply make something (anything) that is compelling, there's often a danger of making a counter-intuitive product or, worse, misleading the map reader through ill-thought-out design solutions.
  Indifference toward the reality in which the map is situated is a failure of the cartographer. Making an unsuccessful map happens when no one understands what the map is saying or they don't particularly care to spend time with it. The form of the map goes a long way to making it work. Asking basic questions about your work-Is it true? Is it relevant? Is it necessary?-is key. Failure to ask these questions is a good way to end up with a dysfunctional map. (Page 134)
• Projections-does the projection support the intended interpretation? For instance, a Web Mercator projection is rarely good for thematic mapping.
  Coloursdoes the use of colour support cognitive processing? For instance, sequential colour schemes do not support our ability to determine quantitative structure in data.
  Thematic techniques-does the technique match the data? For instance, if you have totals, you might use a proportional symbol technique but not a choropleth unless you convert them to a rate.
  3D-does the use of the z-dimension encode anything useful? For instance, 3D is often used gratuitously, and the perspective view creates cognitive issues for map reading.
  Transparency-does the use of transparency introduce visual problems? For instance, transparency is often used on one layer of information to make visible a second layer underneath. This often destroys the form of the upper layer.
  Binning data-does the bin size and shape make sense? For instance, an equal-area projection is fundamental for binning to avoid bins representing unequal areas.
  Symbols-does the symbol design simplify or overload? For instance, symbols often try to do too much. (Page 134)
• The pattern of components on a map is designed to work as a whole to give meaning to the subject matter and communicate information to the reader. The relationships among the various marks are complex, and they can be used support each other to provoke an emotional response in the reader. This emotional response may be passive and simply the response of understanding and interpreting the map. On the other hand, it can be active, and the map can provoke a full range of feelings.
  Clear design is always an objective but we can also imbue maps with powerful imagery. The intent here might be to make a point more forcibly or to give rise to a particular emotional response such as fear, anger, joy, or compassion.
  In just the same way as a picture is worth 1,000 words, the picture itself can be made to relate to the reader on many different levels.
  At their very obvious, propaganda maps have long been used to bring a sense of right and wrong in warfare.
  The graphic elements of propaganda maps are designed to support a very specific response. Certain aspects are exaggerated. Colour is used in a particular way. Perhaps less dramatic, persuasive cartography seeks to be more subtle yet with the same objective. We may see this in certain types of maps used in advertising or journalism where left- or rightleaning political points are being expressed visually.
  Even simple design aspects such as using colour or a type style to bring a particular emotional response can be used to great effect. Red brings out a particular emotional response. Symbols associated with good or bad can be used effectively alongside or as part of the map design. What humans know, see, or experience are all aspects of reality that we can riff off when we design maps. Simple and direct cartography is often a good basic design tenet but adding embellishments that hook into people's views of the world can raise their effectiveness and make them difficult to forget. (Page 152)
• Maps are regarded as being definitive, accurate portrayals of information. That's the power they hold as people examine the lines, colours, wording, and symbology. They develop an impression of trust that is error-free, and this breeds confidence, sometimes over confidence in the efficacy of the product. All maps have errors, biases, and uncertainties.
  Many of these can be related to data accuracy and precision and the way in which these are either dealt with or propagated through the map. Other errors may exist. Biases may be introduced. These may often be unintentional.
  Uncertainty can be a function of the data or the way it is represented. It is the degree to which the way something is mapped varies from its true value. This might be because of measurement errors, the processing and manipulation of the data, or the way in which the cartographer interprets and then represents the data. Error provides a way of measuring uncertainty in a map. For instance, we can quantify positional errors or determine the extent of attribute errors that might be missing or invalid. Errors may be unintentional, random, and difficult to spot though an understanding of potential error can provide a way to assess the extent to which a map can be trusted.
  Bias is an extension of error and reflects a systematic distortion that is introduced into the map intentionally or otherwise. It can be introduced through misuse of data that results in a specific error being propagated across the map, affecting all similar features. It might also be added in more nefarious ways to influence the way in which a particular feature might be represented or a message framed.
  Good editing helps to eliminate a large proportion of error.
  Checking sources, and checking and proofreading the final map remain vital tools. As with most creative works, it's easy to become so close to your own work that you find it difficult to see problems. Have other people look at your map and pass on comment. Often, an obvious error can be easily spotted by fresh eyes. (Page 154)
• All too often maps can be regarded as ineffective and poor because they may fail in many ways. They may be disingenuous about the content; use misguided construction or techniques; or they may even be deliberately designed to be persuasive or propagandist. The ability to accidentally or deliberately make a map in ways that might misconstrue meaning raíses questions of ethics or morals in cartography.
  Maps are made by humans and are therefore inevitably going to contain errors. Errors may result from oversight, poor judgement, or a reliance on software defaults but, sometimes, as a consequence of deliberate action. Mapping is not an exact science; there is no single correct way to map and neither is there a wholly incorrect way (though some techniques can be applied incorrectly). Maps can misrepresent even when the motive is to work toward a map that minimizes the potential for misunderstanding. Here, I've modified Borden Dent's code of ethics for mapmakers:
  • have a straightforward agenda, and purpose;
  • strive to know your audience (the map reader);
  • do not intentionally lie with data;
  • show all relevant data whenever possible;
  • don't discard data because it might be contrary;
  • strive for accurate portrayal of the data;
  • avoid plagiarising; report all data sources;
  • ensure symbols don't bias the interpretation of the map;
  • the map should be able to be repeated by others;
  • be attentive to differing cultural values and principles; and
  • don't let defaults drive your design. (Dent et al. 2008)
  Computers can compound the problems associated with these ethical ideals. Most mapmaking software easily supports basic mapmaking tasks for anyone to make maps.
  This is undoubtedly empowering and inclusive but it can also create issues. Software will, on the whole, provide opportunities but are they appropriate to the original data or the analysis that was intended? The purpose and value of the map should drive the use of the software. (Page 156)
• Flow maps show linear movement between places and can show qualitative difference between types of flow or quantitative movement through magnitude of flow. Lines generally change in width proportional to the quantity they are mapping but changes in hue and value, for instance, can encode multiple variables at the same time or for emphasis.
  The technique supports the mapping of totals with widths of flow scaled proportionally to the values for each segment in the network. Ratios or proportions can also be mapped using this method. If flow is directional, it's usual to incorporate an arrowhead, or some other cap symbol or graphical effect (e.g.
  tapering or gradient-filled transparency) that signifies origin and destination. When lines branch or merge the widths of the smaller lines should equal the width of the aggregated flow line.
  Flow maps tend to be one of four types: network, radial, distributive, or continuous. In addition to mapping quantity of flow, a network's organisational structure is also illustrated.
  Other networks are less defined such as air routes, which tend to show origins and destinations linked by a line, indicative of a route. Radial flow maps (also called 'origin-destination'
  or 'desire' lines) can show links between paired places.
  Distributive flow maps are a modification of the radial flow map that illustrates movement from a single origin to multiple destinations, with the width of the line changing at vertices where it splits to a destination.
  Three-dimensional and animated flow maps can be used to emphasise volumetric information or movement. In 3D, flow can be mapped by using the volume of a cylinder or another 3D shape along lines that represent the three flow types. Another technique that 3D supports is the creation of arcs across a virtual globe such that the linear symbol isn't attached to the ground, but arcs across the space from the origin to destination. Animated flow maps may incorporate a pulsing effect or some other animation that connotes directional flow at speeds that represent velocity (Page 168)
• When we read a passage of text we read in serial. One word comes after another. They form sentences and paragraphs and we build up a sense of what is being said. As human beings we are well drilled in understanding information that is presented to us in a serial, or linear, form. The same cannot usually be said of a map since information is communicated in parallel, and it's up to the reader to disentangle the messages.
  That's why we use graphical principles like figure-ground and visual hierarchy to give the map structure-to try and enforce a sense of linearity in how people read it.
  What we're essentially asking of our map readers is to undertake a multitasking exercise where we bombard them with a lot of parallel information at once. For many people, though, multitasking is not something they do particularly well. Creating complex graphics therefore lessens the potential of a map reader to be able to understand the map. It's a fallacy to assume that putting as much as you can on a map provides more information. In many ways a map is a perfect exposition of the idea that less is more.
  Building focus can be achieved implicitly by omitting all but the essential from the map. In an era of big data when the temptation is to map as much as you can, it's probably more important than ever to reduce the noise on the map. In an explicit way, we can also use graphical devices to give focus.
  One of the obvious devices is simply ensuring the object and core purpose of your map is placed toward the visual centre of the map and to organise other elements around and in support of that core element. Shining a light, literally through graphical effects, to bring focus on a particular element, or to mute others, can also lead readers to the key aspect.
  There are other ways of ordering the map's message. For instance, if you make a web map and use a linear storytelling design, it's perfectly possible to convert a map or a series of maps into a sequence that leads the reader. This idea sits at the intersection of a written passage and a film, effectively using storyboarding to deliver a map-based message. (Page 170)
• Map design requires an appreciation of the basic tenets of geographical inquiry and insight. Geographers ask typical questions such as What? Why? When? But added to these is the crucial Where? question. Where can be quantified by describing position on Earth in terms of coordinates. However, location can be more sophisticated than simply mapping where features exist, and relative location has become key to exploring spatial relationships. For instance, two cities can be spatially described in their absolute sense by defining their latitude and longitude. In a relative sense they may be a certain distance apart. In terms of interconnectedness, the two places may be temporally closer to each other based on different modes of transport. What this means for cartography is that mapping of coordinates; of absolute location, is only one part of the mapping task. Successfully encoding the processes of geographical inquiry allows maps to support insight.
  The sorts of questions we ask of the world, and of data, can be framed around key concepts such as::
  Direction-describes relative location Distance-describes how far away one feature is from another Scale-the size of the area being studied (micro-, meso-, and macro-scale)
  Location with respect to coordinates or relative to other phenomena Areal patterns-describing a distribution over space Spatial clustering-extent of concentration or spatial proximity Incidence and prevalence-the existence of a phenomenon and its relationship (coexistence) with other phenomena Connectivity-describes the linkage between phenomena rather than the phenomenon itself Spatial interaction-characterises the attraction between places and the spread of movement Regionality-exploration of similarities and differences of phenomena in comparison with other places Change-how phenomena differ with the passage of time or under the influence of other phenomena or changing conditions. (Page 238)
• Truth and integrity in cartography are important and showing something on a map brings with it the veil of truth. People believe maps and diagrams as purveyors of accuracy and, therefore, of truth. Mapping the facts should be a fundamental objective. Not lying in the way information is represented is, though, not the same as telling the truth and there can be different shades of truth. Historically, topographic maps have left off features that clearly exist in the landscape such as infrastructure deemed to be official state secrets left off Ordnance Survey maps in Great Britain.
  In thematic mapping the scope to manipulate data or its representation is perhaps easier to accomplish.
  Rational objectives should drive the development of a symbology that fairly represents the data and leaves little margin for misinterpretation. But even when mapped data is accurate, the way in which it is represented can bring different shades to the truth. For instance, a choropleth map almost always requires the data to be classified. Although the data may all be represented on the final map, the way the classification scheme may have been developed will undoubtedly shape the way the information is represented, read, and interpreted. This might be through a firm choice on the part of the cartographer making the map or, simply, through a lack of understanding of the way in which reality can be shaped.
  Choice of classification scheme is only one way in which truth can be manipulated but sometimes the cartographer has very little control. Again, using a choropleth map as an example, we usually map into standard geographies usually enumeration areas that have been used to collect the data. These will almost certainly be predetermined and, consequently, have their own inherent bias as the Modifiable Areal Unit Problem on the opposite page shows. In short, the map is as much a function of the boundaries used as the spatial pattern of the population mapped. If boundaries were drawn differently, data would be collected in different containers and a different visual pattern would likely emerge. (Page 240)
• Data can be measured on nominal, ordinal, interval, and ratio levels, which increase in detail. Measurement assigns numerical values to phenomena to represent certain facts about them, and it is this translation of phenomena into data that ultimately allows it to be mapped.
  Nominal scaling is the simplest level of measurement and is akin to a binary form of measurement where everything can be measured as 1 or 0 (or yes or no). Phenomena are classified into groups that are 'similar in kind' and labelled with identifiers to indicate their difference. Nominal data is not based on any numerical measurement and better described as measuring qualitative, rather than quantitative, difference.
  Nominal data allows the identification of commonality or difference between different phenomena.
  Ordinal scaling measures a rank or hierarchy between phenomena. Phenomena are arranged from least to most or vice versa to discern relative position. There is no attempt to identify the distance between each data value. Some statistical analysis is possible on ranked data, for instance to identify relative positions of ranks between different datasets.
  Interval scaling positions individual events in rank order and identifies the distance between the ranks. Individual data values are numerically scored with equal units used across the scale being measured. In this sense, the interval scale allows an analysis of the magnitude of data values at points along a measured scale. There is no natural starting point (or zero) for an interval scale, and the data is said to have no absolute values; instead data is relative.
  Ratio scaling is similar to interval scaling in that it involves ordering individual events with known distances separating each event. However, the difference on a ratio scale is that magnitudes are absolute because they have a known starting point. The ratio scale of measurement has zero as its starting point. (Page 272)
• Lines on a map have three main functions. Firstly, they define edges and boundaries between adjacent areas that might be demarcated through, perhaps, political determination. They contain a space that shares a particular characteristic. Secondly, they connect one place to another whether that is two features on a map or whether it's a feature and an associated label (perhaps through the use of a leader line). Finally, the line can represent a linear component such as a road, railway, or movement of some phenomena. This final function differs from the first two in that the line itself might represent some two-dimensional character such as direction of travel or magnitude. The line acts as a container for some aspect of the phenomena that it represents.
  With so many functions that a line can perform, the task for the cartographer is to establish symbology that delivers information. Importantly, this should connote the meaning of each of the functions which can be easily understood while not overcomplicating the symbol structure. For instance, it's fairly common on road atlases to demarcate higher class roads by drawing them with two or sometimes three parallel lines. The use of parallel lines is often referred to as casing, as it's effectively a single, often thicker, inner line cased by two outer lines. By definition, this increases the complexity of the symbol and gives the map reader additional cognitive load in unravelling the complexity, but it does make the symbol, and therefore, the feature, seem more important than other lines on the map. A single, thicker line can often perform the same job but with a much cleaner look though the drawback is the line may become too figural and out of visual balance with other map elements.
  It's impossible to get away from the fact that most maps are made up of multiple lines and types of line. Attempting to reduce notational complexity will result in improved efficiency in communication of information. Making lines less sinuous, making them simpler in form, and minimising the type and variation of line work reduces complexity (Page 274)
• In practical terms, which map projection to choose relates to the following questions:
  1. What projection properties must be preserved, as far as possible, for the mapping task, and which are less important?
  2. Are the deformational characteristics acceptable, considering where the mapped area lies in relation to the projection?
  3. Can the projection be manipulated to improve certain characteristics such as recentering it?
  4. Will the shape of the resulting map be familiar to map readers, or does the pattern of the graticule create an overtly awkward appearance?
  5. Is the data to be used in the mapping task in a suitable format for projection or reprojection to the chosen map projection?
  6. Is the projection supported in the software to be used, or will it need to be created? (Page 292)
• Maps are powerful objects and can sometimes become a major factor in life and death. For many maps, the cartographer has little to be concerned about other than trying to be truthful, impartial, and communicate some information. Sometimes a map might offend if it takes a particular editorial view. Rarer, a map may be crucial in decision-making that places human beings in danger.
  Conversely, they may support the preservation of human life, for instance through humanitarian mapping of natural disasters that support emergency planning, evacuation, and resettlement.
  Maps are important tools for military use. Their use in planning as well as ground operations actively supports the fundamental needs of warfare. Historically, maps have been used to plot troop movement and plan out battles. Troops on the battleground must have a basic knowledge of orienteering with maps. Trench maps were crucial in identifying snipers and observation posts. Topographic maps showing gun ranges and compass bearings assisted the identification of targets. In modern drone warfare, digital cartography and accurate positioning are crucial to delivering a drone strike on target. So-called 'friendly fire' and mistaken identities are often a result of poor cartography and targeting as well as misidentification.
  Satellite navigation and digital mapping has also been implicated in deaths. Although it can be amusing to see reports of satellite navigation errors that send trucks down narrow paths, people following turn-by-turn directions and ending up in a river, or people aimlessly walking along major highways, there are more severe consequences. We implicitly trust the databases and digital maps that drive satellite navigation to be accurate but, of course, they contain errors which often place people in danger. Humans are also fallible and can enter the wrong address or location. There have been numerous reported cases of ambulances going to the wrong address, sometimes many miles away from the intended victim who has later died (Page 300)
• Perhaps the biggest impact of web technologies on cartography has been the rise of the map mashup. A mashup is constructed by people taking available data and marrying it with other data or data they provide themselves and then repackaging it. Data often comes in the form of map services. A typical mashup might use a third-party basemap and an operational layer of content. Additional layers such as text or switchable layers of content might also be incorporated.
  Map mashups are a democratised form of mapping. They allow anyone with the basic technical skills to engage in the activity of mapping and this, in turn, leads to a proliferation of maps on all kinds of subjects. Some support citizen science, while some merely overlay the location of houses for sale or some other spatial dataset that can be scraped, or used.
  These sorts of mashups are often characterised by a weak appreciation of cartography since they rely heavily on the way in which a map service has been prerendered.
  Mashups are aggregated. They bring together content to be visualised to reveal new insights based on the precepts of Web 2.0. Web standards introduced the notion of shared common data formats. APIs further support mechanisms that you can use to mix and match different organisational services to develop new services, products, and information.
  Mashups have developed over time. The 'map sandwich'
  metaphor was born out of frustration that many mashups simply don't work because too much information obscures the basemap. Mashups can quickly become unreadable. Rather than having two layers, a map sandwich is built from multiple layers, often placing the operational layer between terrain and an overlaying reference layer (boundaries and names). This is more akin to the compilation of a traditional map but extends the mashup concept. With the advent of customisable vector map tiles, the ability to tailor different layers of a mashup supports more nuanced cartography and supports the ability of mashups to become fully formed cartographic products. (Page 302)
• Maps have become location aware as you move around, and they update to represent features using real-time content that is specific to where you are. It's a paradigm shift from the paper map when you were not at the centre of the sheet. In fact, all too often you found yourself on the edge of a sheet and very often onto a separate sheet. So maps are now edgeless and in some respects scaleless too as you can move swiftly through scales from small to large and back again. Maps now have movement that is tailored to you as an individual user. More than that, cartographers used to map things that didn't change very much. Now, maps show rapid change. Basemaps are updated overnight so when you switch on your maps in the morning they contain the very latest topographic detail.
  The content you see on maps is now heavily derived from sensor networks, live data collection streams, and through wireless transfer. This rapid data collection and update has largely been supported through automating many of the collection and delivery mechanisms for map data and eliminating human input in the sifting and sorting process.
  This has meant maps can move rapidly and has largely solved the age-old problem of map printing, production, and storage.
  This collapse of time has meant mapmaking has gone from years to months to weeks to seconds in only a very short space of time.
  Mapping movement itself has also changed as technologies now support interaction, animation, and immersible 3D.
  Static cartography was always challenged to support people's need to recover information. Pop-ups and other gesture-based map interactions support data discovery and truly exploratory mapping needs. Maps now go beyond showingthey allow people to investigate more deeply and interact with the world as they travel through it. Maps also now include new interface elements, and this has meant that the map itself no longer has to struggle to represent everything or, indeed, to omit detail to preserve clarity. The map might now be seen as the interface to data discovery on the go. (Page 310)