4.2 Fuel Efficiency Now that you have a much cleaner dataset, we can start to look at some of the data more closely for

[answerhappygod]

4 2 Fuel Efficiency Now That You Have A Much Cleaner Dataset We Can Start To Look At Some Of The Data More Closely For 1 (217.68 KiB) Viewed 44 times

4.2 Fuel Efficiency Now that you have a much cleaner dataset, we can start to look at some of the data more closely for insights. In particular, we want to look at the fuel efficiency in litres per 100km. In general, there are many confounding factors and unknown variables that can make an analysis of fuel efficiency difficult: engine size, vehicle type, fuel type (diesel vs petrol will show a massive difference), aircon usage, vehicle load, weather, transmission type. With this in mind, we need to be aware that the results found here are unlikely to be completely representative and accurate, but hopefully indicative. When you start this section, make sure you have removed outliers as indicated in the previous question. 1. Look at the difference in cost per litre per country for January 2022 - use the average currency conversion rate to Rands (quote your values and source). Are there any notable differences? Discuss reasons why this may/may not be the case. [5] 2. Looking at the odometer readings, find examples of where users have missed logging a fill-up. Give a basic rule for identifying this, and estimate how many there are in the dataset. [5] 3. Plot the average distance (in km) per tank per country. Which country has the largest average distance? Provide some explanations for why this might be the case. [5] 4. Do newer vehicles drive further distances between fill-ups? Provide a plot to show this. [4] 5. Take the top 5 most popular vehicles in SA (ie, those with currency set to R). Compute their fuel efficiency and discuss whether these values are realistic. [3] 6. Which vehicles are the most fuel efficient in each country? (Make sure the values are reasonable!!! You can look up values of fuel efficiency online to do a sanity check, but a value of 11 per 100km, or 1001 per 100km are clearly wrong). [5] 7. Plot the difference in fuel efficiency for the top 5 Canadian vehicles between seasons. Would you expect to see big differences, and do you see them? [3] 8. Show the correlations between fuel efficiency and other features. You should find that there is a relative strongly correlation with distance travelled, the age of the vehicle, and the model of vehicle. [5] 9. Use a random forest to get a list of the most important variables. How different are they from each other, and how do these relate to the variables from the correlations above? [5]