Module 10: Dot Mapping

This week was Dot Mapping.  Dot maps indicate locations of an occurrence (in this case conceptual points), where each dot represents a set amount of that occurrence.  Consideration should be taken with enumeration unit (small unit best), dot value (trial & error, nomograph or software) and dot size.  The dots can be placed:  1)uniformly in the enumeration unit (not preferred: impression of continuity across the area), 2)geographically weighted, those of higher values are weighted to be closer to other higher values or 3)geographically based, utilizing ancillary information both limiting attributes and related attributes to place dots based on area attributes.  Data criteria for Dot Density is conceptual data (raw counts), discrete, used to compare or portray variations or patterns. 

There are advantages of Dot Density maps.  Dot Density maps are an easy concept to understand.  They are effective at portraying variations.  Ideally the dots could be counted to recover data.  These maps can be adapted to include other phenomena: urban areas, slope, bodies of water.  Dot Density can be used in correlation with other types of maps. 

There are also disadvantages of Dot Density mapping.  The map can be hard to estimate density.  The map reader could misinterpret a dot as a single occurrence.  Computer dot placement can misrepresent patterns or lack of patterns.  Possible additions to counteract some disadvantages could include a clear legend to avoid the dot as a single occurrence and utilizing ancillary data to avoid random placement.

This week's lab was to represent the 2000 population density of South Florida in a dot map.  Provided in the lab materials were shape files of South Florida, Surface water and urban land use as well as table with census population data.  

The tabular data was joined to the south florida shape file (with both shape file and tabular data (utilizing add data button) displayed in ArcGIS>right click south florida layer>join and relates>join>based on county name, joined based on area).  Dot Density symbology was accessed through the south florida layer properties under the symbology tab (quantities, dot density. population was selected and added as symbol).  Dot size and dot value (1-5) were adjusted until max preview started to coalescence.  For my specific map I utilized a dot size of 4.5 and a dot value of 10,000.  I adjusted the color (several times) to find a color that would stand out and clearly be the highest in the visual hierarchy.  I finally, chose to utilized a bright full purple.  The counties included in the south florida shape file were adjusted and examined with different colors and outlines and without. "Maintain Density by" feature was turned off so the dot value or dot size would not change when zoom in and out.  Properties window within symbology tab was opened and the dots were set to fixed placement, this prevented movement when ArcGIS redrew as changes were made.  Finally, masking was also specified in the properties window.  Initially the mask was set to exclude the surface water layer.  Although this did keep the population points from being located in a body of water, the arrangement of the dots density still had clear differentiation at county boundaries.  A second mask option was applied that placed dots within urban land.  This option presented the dots clustered in and around urban land, more as it would be in the natural world.  The rest of the map was adding essential map elements (title, legend, north arrow, credits, projection,author, date).  Masking the dots really bogs down ArcGIS, so the mask was turned off while final elements were added and style choices made.  Lab instructions requested geographic reference by labeling some major cities in the area.  I choose cities that would be large enough to be recognized.  I added a short integer field to the Major Cities (obtained shape file from previous lab) attribute table named display.  I started editing to input for that field.  I utilized "1" for those cities I wanted displayed and "0" for those I did not want to display.  I stopped editing and saved my edits.  In the symbology tab, category unique feature, value field set to my new field "display", turned off "all other values", "add all value" button, chose symbol for "1" and highlighted the "0" option and clicked remove button, ok.  Due to the placement of the city names amongst the dots I added text for the labels.  I left Tampa's label covered up by dots.  I could not find a place where it would work.  I differentiated the surface water types and added a legend for that layer.  The typical legend by ArcGIS, created for the population layer, only provided the most basic information, one dot = value.  Three visual anchors showing low, medium and high densities. I drew these features in ArcGIS.  I drew a box and copied and pasted it twice to ensure all three where the same size.  I aligned them with the align function.  I added a dot (rectangle drop down box and chose marker), adjust the dot to the same size and color as I utilized in the symbology, and copied and pasted the desired amount.  I symbolized low density with 3 dots (30,000 people), medium density with 20 dots (200,000 people) and high density with 50 dots (500,000 people).  The layer organization I decided to leave the layer with the dot density hollow - without fill and without outline.  I set a second florida shape file as the lowest layer with a bright green for the background.  The surface water layer and major city symbols were in the middle and finally the dot density layer with the urban mask turned back on is the top layer.

Module 9: Flow Mapping

This week the focus was Flow Mapping.  The flow map created in lab this week is a radial flow map, all the spokes (geographic regions) feeding into a centralized hub (U.S.)  There are two other  types of flow maps Parks considered:  a network map shows connectivity of multiple locations (airline route map) and finally distributive map shows flow of  data between geographic regions.  There are two main subcategories of distributive flow maps:  one shows the entire world and attempts to depict actual routes of flow and second depicts flow within a land mass and precision is not as important as general direction and magnitude.  There are also two types of flow maps the text explains that Parks did not consider:  Continuous flow maps, which depict movement of continuous phenomenon such as wind or ocean currents and Telecommunication flow maps, that Parks considered network maps.  The data for this map is quantitative, however, flow maps can be used for qualitative data as well.  The flow lines utilized stylized placement to show spatial interactions as opposed to a specific route from geographic regions to the U.S.  Borden Dent provided essential design decisions in creating flow maps:  flow lines should be depicted as the highest in visual / graphic importance, if flow lines cross smaller flow lines should appear on top of larger, arrow heads are important if flow directions is important, land and water contrast are essential, projection is important, keep it simple, legends clear.  I utilized these guidelines in this map.  The flow lines are the most stylized element in the map keeping it at the top of the visual hierarchy, my flow lines do not cross, arrow heads are included and directional, definite land water contrast, Winkel Tripel projection (compromise projection, relatively minimized distortion of shape, area, distance, and direction, although none of these is really preserved), design is simple, and only the choropleth legend was included as the flow lines are labeled with the actual numeric representation.  The proportional line widths of the flow lines were calculated in excel per the lab instructions.  I chose to indicate that the map was not to scale as opposed to trying to show two scales for seperate elements of the map.  I changed the color of the continental U.S. to white to infer the enlargement of the choropleth in the middle.  I left Alaska in it's location and changed the color to correspond with the data for the choropleth map.  I did enlarge and include Hawaii separately.  I decided to not follow the same orientation for Alaska and Hawaii because Hawaii had to be enlarged to be seen, but Alaska took up too much space if enlarged at the same rate.

Module 8: Isarithmic Mapping

This week the focus is on Isarithmic Maps.  An isarithmic map depict smooth, continuous phenomena.  Isometric maps utilize true point data, data that is measured at a point location.  Isopleth maps utilize conceptual point data, data collected over an area or volume and symbolized as a point (usually the centroid of the area).  This type of map is second most widely used thematic map, behind the choropleth map of last week.  

This week, raster data was provided from USDA Geospatial Gateway.  Data was published by US Dept of Agriculture, Natural Resources Conservation Service, National Geospatial Management Center 09-2012.  Data was originally created by The PRISM Group at Oregon State University. 

I implemented continuous tone by accessing the symbology in the layer properties, selected "precipitation" color ramp. The legend required more work.  Choosing a horizontal style, creating an alternate legend with inverted colors, converting the alternate legend to a graphic, ungrouping and moving the labels.

I utilized the spatial analyst extension (after I remembered to active the toolbar). I used the Int Spatial Analyst Tool to convert the raster values from floating (fractional numbers that have decimal places) to integers, to allow crisp contours.

I implemented hypsometric symbology in the layer properties by opting classified with 10 classes with manual break values (per lab instructions).  Adjusting the manual breaks to whole numbers.  I again chose "precipitation" color ramp, and utilized hillshade relief.

I added contours to the hypsometric tint by utilizing the analyst toolbar this time utilizing the contour list tool.  I included required map elements, followed cartographic design principles.  I added description of the data to the map (after I remember the draw toolbar and the text in shape).  

More from the lecture portion this week:

The fundamental problem in isarithmic map is that data must be interpolated to cover the unknown values between control points, or data collection points.  Methods of interpolation the text discussed for true point data were triangulation, Inverse distance, and Kriging,  Basic (very basic) explanations of each:

• Triangulation: connects neighboring control points to form triangles, utilizes Delaunay triangles similar to Thiessen polygons forms triangles with all points contained in the triangle are closer to that triangles control point than any other. Once triangles are formed contour lines are created by interpolating along the edges of triangles. Finally the contour lines are smoothed.
• Inverse Distance (gridding): layes a grid on top of the control points, estimates values at each grid node, contour points are weighted as inverse function of the distance from grid points.  Consideration is distance only between grid point and control points.  Strategies including search radius for grid points for minimum number of points to fall within radius, and maximum number of whole, quadrants or octants of radius without data are set.  Interpolated contour lines are placed and finally smoothed.
• Kriging (ordinary kriging): used for data without trend or drift.  Similar to inverse distance uses weighted average to compute a value at a grid point.  Unlike inverse distance, consideration is not only the distance from control points to grid points, but also the distances between the control points themselves. More complex method of interpolation, it can produce more accurate map (optimal interpolation). ONLY if one has property specified the semivariograms and associated semivariogram models.  This model also provides a measure of the error associated with estimate, standard error of the estimate, can be established with a confidence interval. 

Symbolization of Isarithmic maps:

• contour lines - lines that mark amounts of a phenomena, can be difficult to visualize
• hypsometric tints - light and dark shades (grey or color) are added between the contour lines to enhance visualization
• continuous-tone - unclassed so color or gray fades and darkens across the entire area to depict different values, can be difficult to get specific data as there are no clear boundaries
• fishnet- gives the effect of a fishnet being draped over the results, can cause blocked information of lower values from higher values.

Module 7: Choropleth Mapping

This weeks assignment we produced a choropleth map showing overall population densities in European countries using 2013 European census data.  Then tied in wine consumption, data from Wine Institute 2012 liter per capita, as a graduated symbol overlay.  In the choice of graduated or proportional, I chose graduated as proportional has too many options for my brain.  Proportional scales the symbol utilized to the data.  Each symbol is proportional to other data, leading to a range of the scale symbol instead of a set number of classes.  I chose a ramp color scheme for the choropleth map to depict the population density, using from forest green to almost white.  I stuck with a green scale to symbolize a ground type color.  The darkest color represents the highest data range.  I chose a more blue green as opposed to a yellow green to help contrast the red for the wine symbol.  The population density legend has no spaces between the individual categories to indicate a continuous range of data.  I chose to utilize natural breaks for the classification method as it seemed to allow for a more natural division as opposed to Equal Interval that forces data into equal divisions or Quantile that forces an equal amount of values into each category.  I eliminated four outliers of small countries with high population density that would have skewed any classification method.  The scale of the map to fit the page would not have clearly shown these four small countries, Monaco, Gibraltar, Malta, and Jersey and with very low wine consumption served more of an element of clutter than information.  I utilized Data Exclusion with a SQL Query to manipulate the data.  I almost utilized a 7 class division as it seemed to allow for even more natural breaks to be represented, however I chose a 5 class division to keep the color distinctions at a maximum.  I labeled the countries in ArcMap and corrected some of the names into english by adding a field to the attribute table with the labels I wanted to utilize.  I further adjusted location of the labels in AI and added a small halo to allow for clearer reading.  The wine symbol started in ArcMap as a half circle in a red wine color.  I added an outline to that symbol as well as added a stem to form a wine glass in Adobe Illustrator.  I chose to keep the stem of the wine glass consistent across the classes to add weight to the bowl of the glass.  The detail of the stem to the wine glass may be too small when the map is considered in the small thumb print for Blogger, but I felt it was adequate to communicate the information.  I added required map elements: legend, scale bar, north arrow, data frame and for this map a small blurb about what the data offers in way of information.