Vehicle Choice Application
Coursera Developing Data Products Assignment
Motivation
- Transportation sector comprises about a third of greenhouse gas emissions in the US.
- Within that, two-thirds of energy use comes from light-duty cars and trucks.
- So, it's very important to know how the consumers make their car purchase decisions.
Greenhouse Gas Emissions in the US (Source: EPA)
'Vehicle Choice' Shiny Application
- This application takes in simple buyer parameters such as driving profile, attitute towards new technology, and so on to see if they can afford zero emissions cars, through a nested-logit approach (that is calculated outside of application).
- An important part of the buyer affordability comes from infrastructure. It could be private (home / work), or it can be public (charging stations, etc.).
- This application can give a guideline on what low emission car technology the buyer can afford over a period of 15 years (2015 to 2030).
- It can also show how the buyer affordability can change based on basic infrastructure vs. public funded good infrastructure.
A demo example
- Suppose if the buyer doesn't drive much (takes the car only to grocery stores), loves any new technology, and plans to buy a car in 2020 (it's a long time, but let's say he finishes his student loan then).
- The model predicts 'Gasoline Hybrid' in a business-as-usual scenario. In this scenario government provides no subsidies for electric cars, and doesn't install any new station infrastructure.
- However, in a good infrastructure scenario, the model predicts 'electric vehicle'! In this scenario, there are a lot of public funded charging stations, which removes the range anxiety of users.
Time series purchase trajectory over time
- The following graph is generated from the table for the buyer profile mentioned in the previous slide.It shows the top 10 technologies feasible for the buyer generated interactively from the R code.
