Quiz
Data Visualization MIDTERM
This quiz has question types that open in the usual player.
Questions in this quiz (51)
1Order the visual encodings (aka mappings) in terms of their effectiveness at allowing a user to perceive a specific value or a close comparison between values. 1: Most Effective, 4: Least Effective
Question 1 options:
[blank_start]3[blank_end]. Area
[blank_start]4[blank_end]. Color
[blank_start]1[blank_end]. Length/distance with a common baseline
[blank_start]2[blank_end]. Length/distance with no common baseline- A3
- B1
- C2
- D4
- E4
- F1
- G2
- H3
- I1
- J2
- K3
- L4
- M2
- N3
- O1
- P4
2According to Weber's Law, perception is not simply on absolute terms, but
- AWe notice only when two large items have a relatively small difference.
- BWe perceive larger differences when comparison stimuli are aligned at one end.
- CWe cannot see small differences between small items.
- DWe perceive differences as a function of the magnitude of the original (i.e. we perceive percentage changes)
3Comparing values between Case 1, Item B and Case 3, Item E requires using which channel?
- APosition
- BPosition with an unaligned baseline
- CCircle position
- DInk density
4What do we mean by pre-attentive attributes?
- Aa fast mechanism for attributing information to visuals before attention kicks in
- Bdata that show on screen before you can pay attention
- Cas single dot of a different color than its background but not with red and green together
- Dvisual cues that are visually apparent very quickly (~200ms) and do not require intentional processing to see
5Gestalt psychology describes
- Ahow the human brain finds structure in visual stimuli
- Bhow color and connection lines work together to override squares
- Chow color and circle size together are recognized in less than 200ms
- Dhow connection between points is only perceived in relative terms
6Based on Steven's Power Law we know the specific coefficients for the linearity of perception of different stimuli like heat, brightness and visual length.
- ATrue
- BFalse
7Which of the following is a reason to use data visualization?
- ATake advantage of human visual processing machinery for insight generation
- BCan get high information density in an image.
- CCorrelation between visual cortex and data values.
- DConveys patterns that may be difficult to describe.
8Which of the following describe(s) a difference between Exploratory and Explanatory visualization?
- AExploratory visualization uses multiple variables simultaneously.
- BExplanatory visualization should be carefully crafted through multiple drafts.
- CExplanatory visualizations cannot include statistical information.
- DExploratory visualizations are usually made quickly and many at a time.
9In data science, visualization is used
- Aonly when machine learning has failed.
- Bby people who have not fully developed a machine learning toolkit.
- Cat the beginning of the process to understand data, in the middle to debug results and at the end to communicate.
- DUntil machines take over the world.
10Which of the following describes the guiding principle for making decisions about how to visualize data?
- AAlways choose the visualization method that matches up with all the variables of the data as well as possible, as long as it is not too complicated for your audience.
- BUse the most sophisticated technique (we'll cover plenty) available that matches the data and is capable of communicating your message.
- CTake into account the data, the people who will see the visualization, and what you want them to understand when they see it (data, audience, message).
11Major differences in the character of datasets can be invisible to summary statistics.
- ATrue
- BFalse
12Tableau allows you to create new visualizations by:
- ADragging and dropping variables onto the axes.
- BSelecting variables and choosing a recommended 'Show Me' graph.
- CBoth.
- DNeither.
13Match the chart type to the data properties.
[blank_start]1. Bar Chart[blank_end] Good for categorical X values and cases where the Y value is ratio scaled.
[blank_start]3. Scatter Plot[blank_end] Can handle multiple Y values per X value
[blank_start]2. Line Graph[blank_end] Implies some importance of the connection between the data points.- A1. Bar Chart
- B2. Line Graph
- C3. Scatter Plot
- D3. Scatter Plot
- E1. Bar Chart
- F2. Line Graph
- G2. Line Graph
- H1. Bar Chart
- I3. Scatter Plot
14Forget about using color or multiple lines or different shapes or other marks or adornment for the moment. In terms of how many data variables can be mapped to the screen, bar charts, line graphs and scatter plots are the same.
- ATrue
- BFalse
15Line graphs show a different type of pattern from bar graphs because
- Ait is easier to read angle (of the lines) then it is to read the position of the dots themselves.
- Bthe lines imply there is a connection between the plotted points, often helping showcase a trend.
- Cline graphs never have ratio scaled variables that need 0 on the Y axis.
- Dlines make more sense interactively.
16As opposed to bar graphs or line graphs, scatter plots
- AA) can have more than one Y value per X value.
- BB) are used to show sampled data (each point is a sample taken of both X and Y axis variables).
- CC) must have 0 in one (but not both) of the axes.
- DA and B
17With both dot charts and bar charts, the axis order is critical.
- ATrue
- BFalse
18Which of the following are true about stacked bar graphs?
- AStacked bar graphs allow you to show information about subgroups of each bar of a bar chart.
- BA proportional stacked bar graph can do the work of several individual pie charts, but may not solve the difficulty reading that same data.
- CStacked bar graphs can be very poor when comparing the size of one subcategory across multiple bars, especially looking for fine distinctions.
- DAll of the above.
19An area chart
- AA) works like a proportional bar graph with connected bars.
- BB) uses the area perception channel and thus requires caution (especially can overemphasize the wrong information, e.g. between bars).
- CC) cannot be used without color.
- DA and B
20In our first tutorial, we looked at a data transformation called melting (or pivoting) that changes the shape of data so that we could map variables to visual features. Which of these describes how that process works.
- AA) The data are converted to a matrix and rotated (transpose) to make the columns represent instances.
- BB) Each data value in the pivoted columns gets its own row in the transformed data.
- CC) A new variable is created that tells which of the original columns a data value came from.
- DB and C
21TreeMaps are an example of which type of visualization
- AA) Bubble chart
- BB) Space-filling visualization
- CC) Hierarchical visualization
- DBoth B and C
22One alternative to a treemap when there hare only a couple levels of depth is to use a (hint: Tableau lets you switch between this easily):
- ADendogram
- BHierarchical bar chart
- CTree Chart
- DHeight map
23A very common example of a hierarchical visualization in everyday computer usage is:
- AFolder browser / File chooser
- BComputer clock
- CPowerPoint
- DIcon tree tool
24Which of these are true about node-link layouts of hierarchical data (multiselect)
- ANodes must not be portrayed as circles.
- BDrawing the deepest part of the tree it is easy to run out of space.
- CThe angle in a circular hierarchical graph is encoding the structure, as opposed to just some data value.
- DUsing a radial layout gives more space for the deeper nodes.
25When hierarchical data is represented in a table:
- AThe depth of the tree is equal to the number of rows.
- BA different categorical column is used to hold the values splitting each layer of the tree.
- CThe data format must include a special linkage between the different cells of data.
- DThe table cannot be long because of the complexity.
26One critical attribute of good data visualization is that data themselves are emphasized. One way to make the data stand out is to:
- AMake the Y axis bounds tighter
- BRemove clutter.
- CUse a wide variety of colors.
- DDraw the data in small, crisp circles.
27One criteria for making sure the data are the most prominent feature of a visualization is the
- ATufte Rank
- BData to ink ratio.
- CR efficiency score
- DRatio of color to lines.
28When graphical integrity is compromised, a graph can be misleading (intentionally or by accident). Watch out for
- AChanging the Y axis to make the data fit tightly.
- BScaling the area used to represent a data point when only the length should be scaled.
- CChanging the X axis so that aggregation in different scales can be misleading.
- DAll of the above.
29The definitions of chart junk and non-data ink cross a little bit, but for our purposes I distinguished between them in class. Match each to its definition.
[blank_start]2. Non-Data Ink[blank_end] "Ink" on a graphic that does not represent data. It may be redundant and removable. Note, ink that represents data should "change as the data change" (Tufte)
[blank_start]1. Chart Junk[blank_end] Visual elements of a visualization, generally graphical elements representing a theme, that detract from the data being shown or distort its representation.- A2. Non-Data Ink
- B1. Chart Junk
- C1. Chart Junk
- D2. Non-Data Ink
30We discussed many guidelines about how to graph information so that the data are emphasized and the message is not misleading. Design considerations can dictate the details of applying these to a particular problem. Which of these principles should guide decisions?
- AMaximizing the amount of the graph's space that is covered by data markings.
- BUse of the latest techniques in data visualization.
- CThe audience and the task they are trying to complete.
- DEnsuring the aspect ratio is equal to 45.
31Which of the following best describes the aspect ratio of a graph?
- Abanking
- Bskew
- Cshape
- Dink density
32One usage of Panel Plots is to show data that would otherwise be a dense set of timeseries lines. Panel Plots
- AA) break the lines out into separate pieces
- BB) can be (unacceptably) confusing/misleading if the axis range are not the same across panels
- CC) are available in Tableau but not R
- DA and B
33Which of the following are types of smoothing used to show patterns in timeseries data
- Amoving average (average of a time window surrounding each point)
- Bweighted moving-average (average of a time window surroudning each point, with more recent points stronger)
- Ctotal average (replace all the points with the average of the entire dataset)
- DLOESS (small regression to see the trend in the immediate area of each point)
34Map each colorspace to its property
[blank_start]2. HSV[blank_end] Moving from the tip of the cone that contains all values of this color space, straight through to the bottom means transitioning from black to white.
[blank_start]3. CIE-Lab[blank_end] Carefully consructed to match human perception of color.
[blank_start]1. RGB[blank_end] Red, Green and Blue are points on a 3D cube.- A2. HSV
- B1. RGB
- C3. CIE-Lab
- D3. CIE-Lab
- E1. RGB
- F2. HSV
- G1. RGB
- H2. HSV
- I3. CIE-Lab
35To create contrast when choosing two colors, pick them
- Awith no more than pi/4 radians difference in their Hue angle (HSV)
- Bwith values 0 and 1 for saturation (in HSV)
- Con opposite faces of the RGB colorspace cube
- Dfrom opposite sides of the color wheel
36Which color is most strongly perceived by your brain?
- ARed
- BGreen
- CBlue
- DIndigo & Violet equally
37Match the type of color mapping to its property
[blank_start]2. Sequential[blank_end] Used for common continuous variables that vary from low to high.
[blank_start]3. Qualitative[blank_end] Maps only to variables of a nominal type, i.e. not continuous. It gets cluttered and confusing to use too big a palette (too many colors) at a time.
[blank_start]1. Divergent[blank_end] Used for variables that have values with two separate extremes and neutral ground in between.- A2. Sequential
- B1. Divergent
- C3. Qualitative
- D3. Qualitative
- E1. Divergent
- F2. Sequential
- G1. Divergent
- H2. Sequential
- I3. Qualitative
38When creating a sequential color palette, pay most attention to varying the:
- AHue
- BSaturation
- CLightness / Value
- DGreen-ness
39Colorblind people
- Asee always in black and white
- Bmay be completely unaware of data distinctions drawn with certain palettes.
- Cdo not know the difference between 74 and 21.
- Dsee always in grayscale.
40To create a palette where there is a large difference between the two ends, use colors close to ____ sides of the color wheel. If there is just a subtle difference from one end of the spectrum to the other, use colors that are close to _____ on the color wheel.
- Areds; blues
- Bopposite; adjacent
- Copposite; opposite
- Dadjacent; extended
41Select all of these that are appropriate uses of color on maps:
- AShow values of continuous data, one carefully colored pixel per location
- BShow different types of terrain, e.g. different color for water vs. trees vs. dessert.
- CEncode topological information, like height above sea level at each pixel
- DAnnotate with categorical color-coded information, like blue lines for cold fronts on weather maps.
42The Earth is flat.
- ATrue
- BFalse
43Which of the following are potential problems using circles on maps to show data at locations?
- AAreas of circles are not a strong data encoding for fine comparison.
- BThe center of a circle is difficult to figure out implicitly, meaning locations cannot be perceived.
- CCommon circle pitfall - must encode area not radius!
- DOverplotting - too many circles makes the map context impossible to read.
44A map where individual area segments are colored to correspond to data values is called a
- AData Mapped Map
- Bchoropleth
- Cmapgram
- Dcartogram
45A map in which the areas of landmasses are morphed to match with some data characteristic is called a
- Aballoon map
- BData Mapped Map
- CTukey Map Plot
- Dcartogram
46A choropleth is also an example of a
- Aheatmap
- Bdata flow map
- Cdensity projection
- Dtreemap
47An exponential function f(x) is one of the form:
- A2*x*x*n
- Ba*x^n
- Ca*n^x
- Db*x^2
48Suppose the values for some variable X are in the range 10 million to 100 million dollars. Using the axis label "Log of Dollars" vs. "Log of Millions of Dollars" means
- AShifting the scale by a factor.
- BShifting the scale by an offset
- CStretching the data
- DStretching only the data above a certain threshold.
49A Tukey Box Plot can be used to show values with uncertainty, i.e. sampled from a distribution. Besides showing the median, it shows context by indicating
- AA) The outliers plotted as points
- BB) One quartile on either side of the median.
- CC) Three standard deviations
- DA and B
50A violin plot with data plotted as dots within the 'violin' area is called
- ACello Plot
- BBean Plot
- CBiHistoGraphical Plot
- DDot-o-lin Plot
51The Quantile-Quantile (Q-Q) Plot
- Ais a special form of bar chart
- Bcan make a comparison between two distributions.
- Cplots the values of two different variables in order.
- Dis functionally equivalent to a Tukey Box Plot
Information
- Description
- Quiz on Data Visualization MIDTERM, created by E I on 02/05/2022.
- Author
- E I
- Language
- English
- Views
- 0
- Rating
- No ratings yet Rate this quiz