本次美国代写主要为R语言数据分析的assignment
Questions:
Electric vehicles are increasingly popular. These cars, however, represent a potential challenge for electric
utilities, because they consume a lot of energy. In order to properly plan for the Grid of the Future, the
Western ILLinois Infrastructure And Management Service’s Electricity Planning Horizon Squad
(WILLIAMS EPHS) need to know how much power a new electric car draws from the grid. They’ve
come to the Harris School to find a team of experts to help them with this problem, but all the faculty are
busy — so you’ve been asked to step in.
1. The EPHS are interested in answering the following question: What is the effect of adding an
electric car to household-level energy usage? In order to get started, they’d like you to present
your ideal experiment. Explain what you would do to answer this question in a completely
unconstrained world, and describe the dataset that you’d like to have to perform the analysis. Use
math, words, and the potential outcomes framework to explain what you would estimate and how
you would do so. Make sure to be clear about the unit of analysis (ie, what is “i” here?)
2. The EPHS, being a budget-conscious agency, can’t provide households with electric vehicles. But
they do have some data that they’d be willing to let you work with. They have a single temporal
snapshot of household energy consumption for their customers. They’d like you to look at
average differences in energy use between households with and without electric cars to get a
sense of what cars do to energy consumption. You have a sense that this is not a great idea.
Describe three concrete examples of why this comparison might not provide the answer that the
EPHS want.
3. Explain to EPHS what the benefit would be of being able to observe households at multiple
points in time. Their database goes from 2007 to 2017, but they only want to share what’s
absolutely necessary, for confidentiality reasons. First, describe, in words and math, what you
would do with data on many households, all of which bought electric cars in 2012. Be sure to
discuss the identifying assumption that would be required for this approach to recover the causal
effect of electric cars on energy use. Provide three concrete examples of concerns with this
approach.
4. Next, EPHS want to know why it would be even better to have data on multiple households,
divided into two groups: households that never bought electric cars, and households that bought
cars in 2012. Explain, in words and math, the estimator that you would use with this dataset. You
should include an estimating equation in the form of a regression. Describe how this larger
dataset would allow you to resolve the concerns you had above. Be sure to discuss the identifying
assumption that would be required for this approach to recover the causal effect of electric cars on
energy use. Provide two examples of remaining concerns, even in this larger dataset.
5. EPHS, given your even-handed discussion of various approaches, are willing to put their faith in
you. They will give you data on the universe of their consumers from 2007 to 2017. This includes
customers who bought cars during many different years. Describe, in words and math, how you
would estimate the effect of electric cars on energy use using this dataset. You should include an
estimating equation in the form of a regression.
6. Use the included ps4_data.csv dataset to implement a simple comparison of average energy
consumption between households with and without electric vehicles. Describe what you find. Use
regression to perform a time-series analysis of the effect of electric vehicles on energy
consumption, using only households who bought electric cars in 2012. Describe what you find.
How does this differ from what you estimated using the initial estimator. Plot energy
consumption against time for households who bought electric cars in 2012. What do you see? (It
may also be helpful to plot average consumption across households). Does this figure affect how
you interpret your estimates? Hint: Net electricity usage might be negative if a household
generates solar power.