upload assignment-2

02_activities/assignments/assignment_2.md

@@ -12,23 +12,31 @@
       ```
       Your answer...
 
+      Good example: https://public.tableau.com/app/profile/oha.center.for.health.statistics/viz/OregonHealthAuthorityCenterforHealthStatisticsWeeklydeathgraph/Dash-weeklydeaths
+
+      The dashboard used for Oregon Health Authority Center for Health Statistics: Weekly deaths graph is an example of a good data visualization because it displays several principles from cognitive psychology/easy cognitive load, perceptual science, and the data is reproducible. The visualization uses a simple line graph with different colours to compare deaths from 2020-22, historical average deaths, and COVID-19 deaths. The concise 2D design allows the audience to easily process and follow the data from each line. It takes minimal cognitive effort to detect patterns, outliers, and directional changes within the line graph. By avoiding 3D geometric designs and multiple colours used in unnecessary places, it minimizes the cognitive load on the audience and increases the perceived factual basis. This allows the users to focus their working memory on interpreting the data. 
+
+      Using consistent colours, spacing, and typographic hierachy, the design choices align with Gestalt principles of similarity, proximity, and continuity. With clearly defined scales and labelled axes, it provides accurate interpretation rather than letting the audience approximate the data. By including clear definitions, notes about methods used, and contextual explanations, it helps the audience interpret the presented knowledge with existing knowledge more easily. In addition, the dashboard has simple filters and hover tooltips that support interpretation and understanding of the data rather than distracting from the results of the visualization. As a result, it has all three qualities of a good data visualization (aesthetic, substantive, and perceptual).
+
+      Lastly, by having the data table/sources, definitions, and data notes available, the visualization is reproducible. The dashboard identifies the dataset (weekly death counts from the state’s vital statistics system) and it provides definitions for key terms such as “week,” “death count,” and any exclusions or reporting rules. This transparency ensures that users understand exactly what is being measured. The visualization also uses standard, simple line charts with clearly labeled axes. So the transformation from raw data to visual form is straightforward and can be replicated in any statistical software. Additionally, the dashboard typically includes updated frequencies, date ranges, and notes about data completeness or reporting delays, which are essential for reproducing the structure of the dataset. Since  the underlying data is publicly accessible through the state’s reporting system, the same dataset can be re-created applying the same definitions, and generate an identical time‑series chart.
 
 
+      Bad Example: https://datavizproject.com/wp-content/uploads/examples/open-uri20140714-23180-qbuljk.gif
 
+      The data visualization for how much space retailers take is an example of bad data visualization because it increases cognitive load, it has limited reproducibility, and it lacks perceived factual basis.
+
+      The visualization increases cognitive load due to the chart type, relational values, many visual elements, and having an exploratory composition. Although the visualization uses a vertical bar graph, it is difficult to recognize the graph type because the design prioritizes decorative elements over clear data encoding. The visualization uses icons, irregular shapes, and area-based graphics. Each bar for the retailer has a different thickness, colour, and length which does not seem relational to the data. The audience has to navigate through the decorative elements first before understanding the data it is presenting. As well, it does not provide a defined scale to compare the data between each retailer. The audience has to make relational inferences about how much area retailers take up instead of having absolute values to look at. As a result, the viewer must expend unnecessary mental effort just to interpret basic comparisons, making the visualization visually appealing but cognitively demanding.
+
+      As well, it is has limited reproducibility because it only provides the citation list. The data visualization does not provide method of collecting the data and access to the raw data to make the visualization. It does not document how the data was cleaned, transformed, or calculations to make the conclusions of the visualization. The visualization relies on aesthetic features rather than quantities features, making it difficult to reproduce the exact quantities. Without the data set, the audience cannot verify the accuracy of the numbers and recreate the visualization. Limited reproducibility hinders the data visualization because reproducibility is used to ensure a trustworthy analysis and validate the analysis. 
 
+      Lastly, the visualization lacks strong perceived factual basis which makes it difficult to interpret and evaluate the results. The design uses an irregular 3D image and irregular geometric shapes rather than a 2D image. Although the overall layout looks clean, the differences between the designs for the retailers' data, labels, and scale values, make it confusing for the audience to interpret the results. Even though citations are provided, the infographic does not show how the data was scaled or transformed, which weakens trust in the factual grounding of the visualization. As a result, the visualization relies heavily on aesthetics rather than substantive and perceptual qualities to present the data which makes it a bad example of data visualization. 
 
 
       ```
     - How could this data visualization have been improved?  
       ```
-      Your answer...
+      This data visualization could be improved by changing the aesthetics of the visualization and implementing measures to increase substantive and perceptual qualities. The visualization should replace the decorative and area-based graphics with simple, 2D visuals. The area-based graphics can be changed to bar charts or proportional bars with a single colour. The insights should have a high contrasting colour so it does not blend into the background of the graph and improve accessibility. This can help ease the cognitive load instead of trying to determine how the area-based graphics and thickness of each graphic related to the total storage acreage. As well, consistent scales and numerical labels should be provided instead of providing the highest number of total stores and highest number of total store acreage. Having consistent scales will help clarify how both of the variables are being measured and help compare the different the retailers. This will help the visualization become more explanatory rather than exploratory. Through consistent typography, spacing, and grouping, the visualization would guide the audience's attention toward the most important insights instead of overwhelming them with competing aesthetics. Lastly, a brief explanation of how the data was processed, the data sources, and methodology should be provided to make the visualization reproducible. However, by providing the data sources, it can also help strengthen the credibility and help viewers understand the factual basis behind the visualization. 
 
-
-
-
-
-
-
       ```
 - Word count should not exceed (as a maximum) 500 words for each visualization (i.e. 
 300 words for your good example and 500 for your bad example)