Part 1

Here is a prompt that includes a brief introduction to the U.S. National Oceanic and Atmospheric Administration’s (NOAA) storm database.

Assume two data analysts have been given the same prompt above and you are the audience who will evaluate the two analyses. Assume you do not have access to the database itself until you see the data analyses. Create a rubric that you would use to evaluate the data analyses.

Helpful hints:

The rubric needs to include:

• A bulleted list of objective characteristics (or data analysis principles) that are important to you as the audience. (Note: you can include characteristics outside the six principles that we discussed in class if you believe that they are objective characteristics about data analyses). The total number of data analysis principles (\(N\)) can be as long as you wish, but it must include at least four (i.e. \(N \geq 4\)).

For each data analysis principle (\(i\)), include:

• 1-2 sentences describing why you are interested in these principles as an audience. If a listed principle is outside the six principles discussed in class, you also need to describe why you think they are objective characteristics.
• A weight \(w_i \in [0, 1]\) that represents how important this principle is in your rubric. The weights for each principle should sum to 1 (i.e. \(\sum_i w_i = 1\)). For example, if you include reproducibility in your rubric and reproduciblity is not very important to you, you might assign it a small weight.

Note: You could just exclude reproduciblity entirely from your rubric (e.g. this is a principle that you as an audience are ambivalent about and it does not factor into your evaluation of the data analyses). Alternatively, you could include it in your rubric, but intentionally downweight the importance of reproducibility by giving it a small weight (e.g. maybe analyzing the data as fast as possible is your priority and you do not want the analyst wasting time on making it reproducible). These are two different approaches that you should carefully think about when building your rubric.

Note: Part 1 is due Friday September 13, 2019 at 11:59 PM. Write down your answer in this R Markdown document, knit the R Markdown, commit your changes, and push to GitHub, but you also need to submit your answer to this google form by the due date. We will release Part 2 on Saturday September 14, 2019 at 12:01 AM.

Part 2

On Saturday September 14, 2019 at 12:01 AM we will release the links to the two data analyses by sending the links on Slack.

rubric <- data.frame(data_analysis = rep(paste("analysis", 1:2, sep="_"), each = 2),
                     principle = rep(c("reproducbility", "exhaustive"),2),
                     weight = rep(c(0.1, 0.9), 2),
                     score = c(10, 4, 2, 6))
rubric

##   data_analysis      principle weight score
## 1    analysis_1 reproducbility    0.1    10
## 2    analysis_1     exhaustive    0.9     4
## 3    analysis_2 reproducbility    0.1     2
## 4    analysis_2     exhaustive    0.9     6

Part 2.2

In this part, answer the following questions:

1. Now that you have evaluated the analyses with your rubric, in your opinion, what problem do you think the data analysts who developed the two analyses are addressing or solving? (Note: they could be different problems / goals for each analysis).
2. What type of question is being asked (e.g. predictive, causal, inference)? What evidence do you have to support this?
   
3. What audience do you think the two data analysts built the analyses for?
4. Using the principles in your rubic, compare and contrast the differences between the two analyses. Describe in words what are the similarities or differences. Give examples of data analytic elements that are used in the data analyses to support your argument.

Part 2.3

In this part, we will now introduce two different audiences. Instead of assuming you are the audience, let’s assume there are two new (but different) audiences:

1. The head of the United States Environmental Protection Agency (EPA). This is an independent agency of the United States federal government with a goal of protecting the environment.

2. A potential homebuyer who is interested in purchasing a house along the coast in the state of Florida.

Reflecting on these new audiences:

1. Modify the rubric to reflect the principles and weights that these audiences might be interested in. We know you are unlikely to be either of these audiences, but use your best judgement to make an educated guess. The principles themselves and the number of principles can change between audiences, but again there must be at least \(N \geq 4\) for each audience.
2. For each principle, give 1-2 sentences describing why these audiences might be care about these same (or different?) principles that you have now included for each audience.
3. Give a score for each principle you include in the rubric for each data analysis and audience.
4. Once you have completed this, think about and make specific suggestions on how could the two analyses could be improved (e.g. what changes could be made to make it better from the perspective of the new audience members)? Keep in mind, you might make different suggestions for each of the two audiences.

Update the data frame you created above to include this new information. Include a new column called audience corresponding to which audience is evaluating the data analyses. For example, if there were only two principles included in rubric that all three audiences cared about, this is an example of what the data frame might look like

final_rubric <- data.frame(data_analysis = rep(paste("analysis", 1:2, sep="_"), each = 6),
                     audience = rep(rep(c("me", "EPA", "homebuyer"), each = 2), 2),
                     principle = rep(c("reproducbility", "exhaustive"),6),
                     weight = rep(c(0.1, 0.9), 6),
                     score = c(10, 4, 2, 6, 1, 9, 4, 3, 4, 5, 8, 4))
final_rubric

##    data_analysis  audience      principle weight score
## 1     analysis_1        me reproducbility    0.1    10
## 2     analysis_1        me     exhaustive    0.9     4
## 3     analysis_1       EPA reproducbility    0.1     2
## 4     analysis_1       EPA     exhaustive    0.9     6
## 5     analysis_1 homebuyer reproducbility    0.1     1
## 6     analysis_1 homebuyer     exhaustive    0.9     9
## 7     analysis_2        me reproducbility    0.1     4
## 8     analysis_2        me     exhaustive    0.9     3
## 9     analysis_2       EPA reproducbility    0.1     4
## 10    analysis_2       EPA     exhaustive    0.9     5
## 11    analysis_2 homebuyer reproducbility    0.1     8
## 12    analysis_2 homebuyer     exhaustive    0.9     4

Save this data frame to .csv file in your github repo with the title 711-HW1-scores-<yourlastname>.csv where you replace <yourlastname> with your last name.

readr::write_csv(final_rubric, path = "711-HW1-scores-<yourlastname>.csv")