Misc. Information

Part 1: Visualization Critique

MIGRATION

  1. Who is the intended audience? The intended audience is people who are interested in migration and population trends in the U.S. This could range from real estate agents, to government agencies, to prospective homeowners, to curious citizens who just want to know a little bit more about U.S. migration patterns.
  2. What tasks does it enable? This collection of visualization enables the audience to gain general knowledge about where people are moving to, where people are moving from, and which states are popular places to live in the U.S.
  3. List all the data types this visualization represents (quantitative, ordinal, nominal). The following data types are included in this visualization:
    • Geographic location (both quantitative and nominal; quantitative in the sense of physical longitude-latitude locations, nominal in the sense of the artificial grouping into "regions")
    • Number of people who moved to and from regions in 2007 (quantitative)
    • States (nominal)
    • State population changes (quantitative)
    • Region populations in a particular year (quantitative)
    • Year (quantitative)
    • Percentage of population born in another state (quantitative)
    • Percentage of population born in state and still living there (quantitative)
  4. How is each data type visually encoded? The following lists the encoding for each data type:
    • Geographic location - physical location is encoded with a map of the U.S., region is encoded by strings representing the region name ("Northeast", etc.) and geographic boundaries drawn on the map
    • Number of people who moved to and from regions - encoded both as numbers and as the lengths of bars in a bar graph. A different color (red) is used to encode the "To South" data, while grey is used for the rest of the presented data of this type
    • States - encoded as strings ("Arizona", etc.)
    • State population changes - encoded as numbers in a table ("+3.48 million", etc.)
    • Region populations in a particular year - encoded as different fill colors for regions on a geographic map ("21%-25% as red", etc.)
    • Year - each year's data is encoded in a different "small multiples", i.e. as different maps of the U.S.
    • Percentage of population born in another state - encoded both as numbers and as lengths of bars in a bar graph
    • Percentage of population born in state and still living there - encoded both as numbers and as lengths of bars in a bar graph
  5. What design principles best describe why it is good / bad? What perceptual principles and color design rules are followed / violated?
      For the "Going South, 2007" visualization:
    • The data is not presented in a clear, easy to understand way. You have to first scan the right side of the bars to figure out the "from" location and then move your eye back left to find the "to" location and finally scan to the right again to get the number.
    • It is hard to see the stories presented by this visualization. It is hard to infer general migration trends due to this bar graph encoding.
    • Two dimensions of data ("To" and "From") are encoded on a one-dimensional scale, which partly explains why it is hard to read.
    • There is bias, in that it does not show the complete picture. How many people moved from to the Northeast from the West? Such data is not shown.
    • The creator of this visualization obviously wished to show that the most people are moving to the South, but shows inherent bias by coloring the bars for "To South" with a bright red color, and only using grey for the other bars. This pertains to rules R1, R2, and R11 in that it incorrectly uses a vivid, high saturation color to excite emotions in a biased way.
      For the "Population change for select states" visualization:
    • Using only a table for this visualization makes it harder for the audience to infer trends. A simple visualization, such as a bar graph, would clearly and cleanly show the trends. Nevertheless, an advantage of this visualization is that it is clean and simple as it is.
    • There might be bias in the choice of the number "12" in choosing to show only the 12 highest ranked states. It is possible that states 13-20 are Northeastern states, which could refute the trend that the creator is trying to show.
    • Without showing the percentage population increases in addition to the absolute population increases, this data could be misleading, since population patterns may be correlated with population sizes in addition to state locations.
    • Unclear frame of reference. What is the "change" over, one year or ten years?
      For the "Population Distribution" visualization:
    • As described above, the use of red in this visualization is both distracting and misleading, in accordance with color rules R1, R2, and R11.
    • The shades of grey used for the encoding of population are hard to distinguish. The creator should follow color rule R4 to create a better color encoding.
    • Despite the color issues, the use of small multiples in this visualization is well-employed, and makes it easy to see changes in the overall distribution.
    • Coloring regions uniformly could be misleading. For example, it is possible that the population is very high in California but relatively low in Oregon, which cannot be seen if the entire "West" region is colored uniformly. Coloring states instead of regions, or using a continuous color scheme in a heat-map fashion would more accurately present the data.
      For the "Magnet States" and "Sticky States" visualizations:
    • The red color is distracting (as per R2), but is acceptable since all the bars are colored the same.
    • That these two visualizations are presented in the same way makes it easier to read.
    • It would be helpful for the creator to also include a bar representing the average in both statistics to provide a frame of reference. Otherwise, these statistics could be misleading.
  6. Why do you like / dislike this visualization? Overall, I like the visualizations in this part because they are simple and easy to read, especially with the use of small multiples and bar graphs. But the use of red here (as well as in the other two parts as well) is unpleasant. I would have appreciated a color scale that is cooler and easier to distinguish. I also like that these visualizations are thematically linked, presenting the overall migration trend towards the South region of the U.S. But is also a problem in my eyes, because the creator clearly crafted these visualizations with that intention, and in such a way that the visualizations are slightly biased. I would have liked a more unbiased presentation of this trend.

MAJOR METROS

  1. Who is the intended audience? The intended audience is people who are interested in the characteristics of U.S. metro areas, specifically about industry/job statistics in those areas.
  2. What tasks does it enable? The visualization enables the audience to understand the dominance that the metro areas have in most major industries, especially when compared to how little land area these metro areas actually encompass.
  3. List all the data types this visualization represents (quantitative, ordinal, nominal). The following data types are included in this visualization:
    • Economic sectors (nominal)
    • Industries (nominal)
    • Land area of metro areas (quantitative)
    • Metro area population as a percentage of U.S. population (quantitative)
    • Percentage of U.S. "x" located in metro areas, for x in {research universities, jobs, graduate-degree holders, knowledge-economy jobs, patents, air-cargo, R&D employment, air-passenger boardings, venture-capital funding, public-transit passenger miles} (quantitative)
    • City names (nominal)
    • City locations (quantitative)
    • Share of industries in metro area (quantitative)
  4. How is each data type visually encoded? The following lists the encoding for each data type:
    • Economic sectors - encoded as strings and colors ("Innovation as light grey", etc.)
    • Industries - encoded as strings ("Biomedical", etc.)
    • Land area of metro areas - encoded as a number ("12%") and as the area of the small red circle
    • Metro area population as a percentage of U.S. population - encoded as a number ("65") and as the area of a circle
    • Percentage of U.S. "x" located in metro areas, for x in {research universities, jobs, graduate-degree holders, knowledge-economy jobs, patents, air-cargo, R&D employment, air-passenger boardings, venture-capital funding, public-transit passenger miles} - encoded as numbers and as the areas of circles
    • City names - encoded as strings ("Seattle", etc.)
    • City locations - encoded as locations on a U.S. map
    • Share of industries in metro area - encoded as numbers
  5. What design principles best describe why it is good / bad? What perceptual principles and color design rules are followed / violated?
      For "The urban share of the American economy" visualization
    • Encoding the percentages as circle area is a poor choice. We know that humans are bad at judging distances in areas, especially circular areas.
    • On top of the poor choice of encoding, they chose to overlap the circles, which makes it even more difficult to judge differences in circle area.
    • The nominal encoding of economic sectors is difficult to differentiate. The shades of grey are not easily distinguishable in print. In accordance with R4, they should also be changing the hue for each encoding.
    • Once again, the use of red for only one data point (land area) is distracting.
      For the "Major metro areas with the largest share of select industries" visualization:
    • The map here is unnecessary, and in my opinion counts as chart junk. Showing the geographic locations of these cities does not add any insight into the trends shown in this part.
    • The charts on either side of the map would be much easier to read if the city names were placed next to each industry.
  6. Why do you like / dislike this visualization? I really dislike this part of the visualization. The circle encoding is very difficult to read, and the map on the bottom does not add anything to the information shown. But more generally, this section does not add to the main purpose of this page in Newsweek, which appears to be trying to show migration patterns. This visualization certainly expresses the disproportionate share of jobs and industries in metro areas, but does not add any insight to migration patterns in the U.S.

URBAN SPRAWL

  1. Who is the intended audience? The intended audience is people who are interested in the living and migration patterns in and around cities.
  2. What tasks does it enable? It enables the audience to see how the makeup of the metro population has changed overtime, specifically that the percentage of the metro population that lives in the suburbs has increased over time.
  3. List all the data types this visualization represents (quantitative, ordinal, nominal). The following data types are included in this visualization:
    • Living area classification ("Suburban", etc.) (nominal)
    • Share of metro population living in particular areas for different years (quantitative)
    • Year (quantitative)
  4. How is each data type visually encoded? The following lists the encoding for each data type:
    • Living area classification ("Suburban", etc.) - encoded as strings, and also encoded pictorially as rings of a circle
    • Share of metro population living in particular areas for different years - encoded as numbers and as wedges in pie graphs
    • Year - each year's data is encoded as a different "small multiple", i.e. as different pie charts
  5. What design principles best describe why it is good / bad? What perceptual principles and color design rules are followed / violated?
    • The "city map" is cute but not very useful. Once again, the red in this color scheme is distracting. A simple list of definitions would suffice for this small visualization.
    • The "small multiples" pie charts are great and are very clear. My only issue is again the choice of red, which exaggerates the "Suburban" wedge (an effect that was likely intended by the creator). Also, the light shades of grey make the already-hard-to-see "Other" and "Exurban" wedges even more difficult to distinguish.
    • I think the main point of this visualization, to show the increase in suburban population, is not emphasized. The addition of a bar graph or scatter plot would make these changes over a time very easy to see.
  6. Why do you like / dislike this visualization? Of the three parts, I like this visualization the best because it makes good use of small multiples and employs very simple visualization techniques. But it needs a good bar graph or scatter plot to show the changes in population percentages over time.