associations

  • combining data on the map can be used to look at association between dataset and location
    • exposure and disease: combined polygon data and points with residential locations
    • crop yields and soil type: polygon with polygon data combination

overlay analysis

  • data is often combined using overlay analysis
  • stacking spatial data on top of each other and processing a combination of both spatial features and attributes to create a new layer
    • overlay multiple data layers
    • modify geometry
    • create new information

vector overlay

  • estimates the geometric intersection of multiple datasets of the same or different geometries and combines them to create a new output dataset

boolean logical operators

  • overlay analysis operations are analogous to boolean logical operators
    • AND: intersection overlay
    • OR: union overlay
    • NOT: area exclusive to only one type

analysis framework

  • data from different sources are collected and laid out together in the same framework using spatial analysis

result variability

  • take care of the changes that may occur when combining data
    • changes maybe made to the data during analysis
  • boundaries of the analysis area determine the aggregated values
    • the analysis for the same location may yield different results
    • if different area boundaries are used
  • same issue applies to scale of analysis
    • results are function of area used to analyze the data
    • take care when visualizing a map with area data
  • this variability is addressed by normalization
    • totals should be normalized when being mapped by area
    • this converts counts to intensity
normalization: done to compare data from different scales and different boundaries
  • ratios such as means, medians, percentages and rates are already intensities (normalized)

choropleth map

  • a thematic map in which areas are distinctly colored or shaded to represent classed values of a particular phenomenon
  • always should show normalized data, not counts
    • since they are mapped over unequal areas or populations
    • this is very commonly overlooked and is a common error

cartography

  • maps are essential to communicate result obtained from spatial analysis
  • technology makes it easy to make great maps
    • at the same time, it makes it easy to make really poor ones to
    • the poor ones are usually the ones that overkill technology use
      • that distort info to be conveyed by the map
      • without understand how the distortion affects perception
      • of the information they are trying to convey
  • tools in the GIS software exist to make good and well-meaning maps
    • go beyond the tool defaults
    • the defaults rarely are exactly what you need to convey the true meaning

carto-fails

avoid rainbow colors
  • use monochrome - light to dark
  • to convey more or less
avoid choropleth map totals
  • using count values for unequal areas and unequal populations
  • to show the counts
  • this is like comparing apples to oranges to draw conclusions
  • normalize with area to show rate/density/intensity in choropleths
  • count vs per capita
use symbol maps with size of symbol indicating count
  • to show counts
  • instead of choropleth maps
projection selection is important
  • understand limits of each type of projection
  • state the projection used as a disclaimer if you are unsure how the visual data is distorted by the projection used
3D maps fails
  • do not use projection view for 3D extrusion visuals
    • projection views look ‘cool’ but have not information conveying function
    • unless your intent is to create a fancy web page banner
  • use isometric views to convey real meaning
    • don’t let fancy get in the way of conveying the core meaning
    • this is cool for the person who benefits from the info provided
  • make interactive 3D maps and add rotation to ovoid occlusion
    • helps retain a coolness factor while keeping the info to be conveyed intact
transparency overkill
  • mashing up thematic schemes on a basemap is common
    • thematic coloring of a basemap is introduces a lot of visual clutter
    • makes the information more cluttered than necessary
  • usually done when people are trying to show off being able to use the GIS tool over using the tool to convey real meaning
    • defeats the purpose and looks tacky and unprofessional to the trained
  • avoid basemaps altogether, especially if you are using cholopleths
    • just the cholopleth is 6 colors
    • but the basemap with the choropleth is 44000 colors
    • it is practically useless to gather any useful data
    • (you can claim time spent on the project though, so some people will do it anyway, it looks like a lot of work was done)
  • if you have to choose a basemap, pick something that is plaid and dull
    • so it is un-obtrusive with the choropleth color scheme
death by push pin
  • you can kill the information that is to be conveyed by adding a thousand pushpins on the entire map
    • couple it with a rainbow colored cholopleth and multi sized pushpins, and you might as well have your customer do coke to take them on a trip
    • to avoid this mess, show restraint on the use of pushpins
    • simple map: easier to quickly digest information
  • convey the meaning with out destroying the meaning with the tools used to convey the meaning
    • gridded choropleth in place of randomly scattered symbol map
    • the data is there, but only for the sake of just being there
    • nothing useful is conveyed
  • analysis paired with sensible cartography supports sensible information interpretation
overstating facts
  • the map is only as good as the data and the analysis put into it
    • be careful of evaluating the value of information conveyed by the map
    • it is easy to be carried away and over-state the info
  • if you underestimate the bias, uncertainty and error of the data source, and make a map and sell it because it looks ‘cool’, it doesn’t aid any meaningful interpretation of what is in reality
    • even simple layer ordering can change the entire meaning conveyed by the map
    • even though all the data ‘is still there’
  • maps made to look cool lie and maybe considered data art, but it is not analysis or cartography
    • maps lie
    • all maps can lie
    • the goal of analysis and cartography is to extract the truth and present it to the viewer
    • data art is not analysis or cartography

evaluating maps

  • it is great to be inspired by maps made by others
  • but if you don’t know what you are looking for, it is hard to decide what is good or bad

resources to make better maps