Cyber Policy: Problem Set 1
About the Problem Set
This problem set covers the technical details on data management, statistical inference, causal inference, and research designs. It is designed not only to check knowledge of the subject but to help better understand key concepts we need to understand the cyber policy.
Getting started
1 Informative summaries of sampled data
In the first three sections of the problem set, we will use data on ransomware attacks that we already discussed in the class. Our key outcome of interest (Y) is the amount of requested ransom. We start by creating informative summaries of the sampled data.
Find and store minimum values of the ransom requests in d in minimum.ransom.
Isn’t it a strange value? Can you think of reasons we have a ransom request of negative value in d?
Create a histogram of distribution of ransom using ggplot(). Add position = ‘identity’ within geom_hisotgram and alpha = .3 to make it more illustrative.
Recreate the plot, but make a separate histogram for each size of the company. All three histograms should be on the same plot.
Recreate the plot, but make a separate histogram according to the indicator that a company is a part of a critical industry (critical_industry). Both histograms should be on the same plot.
Which one of these two variables seems to be more informative for understanding (or predicting) the amount of requested ransom? Explain your rationale in plain English:
Create an object n.countries that will store the count of unique countries in d
Create an object types.of.attack that stores top 5 of types of intrusions (variable type in d)
What can you say about the capacity of companies to classify the attacks they face?
What is the percent of companies that used encryption on their servers before or during the attack (no_encryption)?
