|
|
| Line 1: |
Line 1: |
| == Assignment description ==
| | http://www.clomidlmt.com/ online clomid - buy clomid online wthout prescription |
| [http://ieg.ifs.tuwien.ac.at/~aigner/teaching/infovis_ue/infovis_ue_aufgabe3.html Description of the third task]
| |
| === Assigned Graphic ===
| |
| [[Image:Radon_Level.jpg]]
| |
| Established Percentage of Homes that Exceed EPA's Recommended Level for Radon
| |
| | |
| == Terminology ==
| |
| Edward Tufte introduced a concept that he coined the '''"data-ink ratio"''' [Tufte et al., 1999]: The data-ink ratio describes how much of a graphic represents salient data. The equation according to Tufte is as follows: '' (data-ink) / (all-ink used to print the graphic) = data-ink-ratio''. The data-ink ratio resides in the interval 0 and 1. The goal is to maximize the ratio by eliminating unnecessary and redundant data. <br>
| |
| {{Quotation|In anything at all, perfection is finally attained not when there is no longer anything to add, but when there is no longer anything to take away.|Antoine de Saint Exupery}}
| |
| | |
| == Critics ==
| |
| | |
| === At First Glance ===
| |
| One of the first things that the observer sees by taking a look at the graphic is, that nearly a half of all the US states cannot be matched according to the legend. For the blank (white filled) states there is no further explanation in the legend. So, it's not clear if the states haven't been considered in this graphical representation of statistical data or if it's simply the result of a bad copy or print.<br><br>
| |
| {{Quotation|Graphics must not quote data out of context.|[Tufte et al., 1999]}}
| |
| | |
| ===Colors And Fill Styles===
| |
| Because of the different font sizes of the abbreviations we assume that states with higher percentages are those that are more important to us, because they are at higher risk than the others (amount of nuclear radiation). The use of different font sizes in this graphic is not the ideal solution. A better way would be to use more saturated colors or gray values for those states we want to highlight. <br><br>
| |
| | |
| It's difficult to detect differences in the shading of two used gray values (0-10% and 25%+). Using colors or gray values that visualy seem to be different from each other would make the states more distinguishable in this graphic - concerning the fact that different statements (percentages of homes) should be related to the states. <br><br>
| |
| | |
| Additionally, for the purpose of pointing out the main message of the representation - "some countries have higher percentages than others, hence, there is more radiation and more risk" - it's very useful to use distinct color or gray values to emphasize and highlight exactly this fact. For other graphical representations it's often better to use similar color values. <br><br>
| |
| | |
| Of course, sometimes it is not possible to use colors for such graphics because of economic issues, as common daily print media for instance.<br><br>
| |
| | |
| There is a variety of fill styles (solid monochromatic fills and patterns) in the graphical representation that couldn't be easily interpreted by the observer. Each percentage range is either represented by a pattern <b>or</b> a solid fill color. It is not very clear, why some ranges/classes are visualised by patterns and others not. The class with the lowest percentage (0 - < 10%) has a striped fill pattern. The next class (10 - < 15%) has a solid color fill that is nearly black. The following class (15 - < 20%) has a slightly lighter fill color than the previous one. The next one (20 - < 25%) uses a pattern again whereas the following is filled with a solid color again. It would be better to choose one fill style - solid color or pattern. Otherwise an observer could misconceive the graphic as interpreting each fill style - solid color or pattern - belonging together. And this would be a wrong message given by the representation.
| |
| | |
| ===Data Ink (Ratio)===
| |
| ====Lines And Borders====
| |
| For the purpose of visualising the borders of the US states, relatively thick lines were used. Using thinner lines would save some data-ink and would perhaps improve a bit the reading efficiency. <br><br>
| |
| | |
| The exactness of the map is another point we want to discuss. Some US states in the map have a double line as state border (such as the borders between North and South Dakota; Wisconsin, Iowa and Illinois; Kansas and Oklahoma and some others). Another defect is that some states seem to be drawn upon another state because of bad drawn borders.
| |
| The missing accuracy in this graphic leads to confusion of the observer. Such defects should be avoided.
| |
| | |
| ====Written Data / Labeling====
| |
| The use of larger font sizes for the abbreviations of the states according to the class they belong - small font sizes for low values and larger font sizes for high percentage values - is a valid way (according to [Few, 2004a]) to point out some data, but in this case, due to the amount of data and the usage of different gray values, it's not clear why the font sizes vary (at first view). It looks a bit unorganized. <br> For instance, the area covered by Michigan (MI), Indiana (IN), Kentucky (KY)and Tennessee (TN): At first glance, this states seem to share some information or data because the font-size used is the same for all of these states. But in fact, Michigan is the state with the lowest percentage value by comparison with the other mentioned states. The use of one distinct font size and a simple color or fill pattern for differentiation would be more than enough.<br><br>
| |
| | |
| A problem is the labeling of very small states, e.g. states like Maryland (MA) which have it's location in the nothern part of the east coast. The placement of labels in the center of states is not possible here. Choosing smaller font-sizes is not the best solution because the letters would not be readable any more. A solution could be to enlarge the whole graphic or to put - like it is realised in the graphic - some arrows or lines which point to the states.<br><br>
| |
| | |
| The use of state abbreviations instead of full names as labels is the right choice here because there would be hardly any place for long names. But in addition, for the sake of clarity it would be nice if a legend could explain the abbreviations by giving the full names of the US states - for people who do not know the US state abbreviations.<br><br>
| |
| | |
| == Improvements ==
| |
| *First thing we considered to be changed in the original graphic were the fill colors / patterns. In the "original" image, these are not consistent because solid colors and patterns were mixed up. We decided only to use solid color fills for different classes. States that are at higher risk are darker than the states in lower percentage ranges. <br><br>
| |
| | |
| *We use the same font and size for all labels to make the graphic better legible. Hence, only the color shading visualises which state is member of which class.<br><br>
| |
| | |
| *The thickness of the borders between the states are thinner than in the "original" image. In the legend at the bottom of the graphic the data is ordered by the percentage ranges (an therefore also by the color intensity). This makes the graphic easier to understand.<br><br>
| |
| | |
| *We have tried to follow the data-ink ratio model and to eliminate all unnecessary visual elements from the original graphic in order to make it much clearer - we draw the background white and lines a bit thinner. <br>
| |
| | |
| [[Image:Usa_Kopie.gif]]
| |
| | |
| == Comments ==
| |
| "We have tried to follow the data-ink ratio model and to eliminate all unnecessary visual elements from the original graphic in order to make it much clearer."
| |
| | |
| ''Soll heißen: Ihr habt den grauen Bereich um die Karte entfernt, oder?''
| |
| | |
| A: Ja, wir haben den grauen Bereich entfernt und die Linien etwas dünner gezeichnet.
| |
| | |
| ''Sonstiges:''<br>
| |
| ''- keine exakte Klassenbildung (besser: 0 - < 10%, 10 - < 20%, usw.)''<br>
| |
| | |
| A: Haben wir nun geändert.
| |
| | |
| ''- Die Grüntöne sind mit Ausnahme von "above 25%" z.T. sehr schwer zu unterscheiden (insbesondere 15-20 und 20-25).''<br>
| |
| ''Ich würde hier eher eine Skala von Hellgelb über Orange bis Dunkelrot verwenden (zudem wird Rot eher mit Gefahr assoziiert als Grün).'' <br>
| |
| ''Für "Undefined" würde ich ein Hellgrau verwenden, das wäre neutral.''
| |
| | |
| A. Haben wir ausgebessert.
| |
| <br>
| |
| <br/>
| |
| | |
| == Links ==
| |
| | |
| | |
| * [[Teaching:TUW_-_UE_InfoVis_WS_2007/08|InfoVis:Wiki UE Homepage]]
| |
| | |
| * [http://ieg.ifs.tuwien.ac.at/~aigner/teaching/infovis_ue/index.html UE InfoVis]
| |
| | |
| *[[Teaching:TUW - UE InfoVis WS 2007/08 - Gruppe 07|Gruppe 07]]
| |
| | |
| ==References==
| |
| *[Few, 2004a] Elegance Through Simplicity. Retrieved at: December 09, 2007. http://www.intelligententerprise.com/showArticle.jhtml
| |
| *[Few, 2004b] Show Me the Numbers: Designing Tables and Graphs to Enlighten, Analytics Press, 2004, Chapter 7 - General Design for Communication.
| |
| *[Tufte et al., 1999] Tufte Design Principle, Retrieved at: December 09, 2007. http://ldt.stanford.edu/ldt1999/Students/mizuno/Portfolio/Work/reports/tufte/ed229c-tufte-outline.html
| |