EECS 298: Social Consequences of Computing

Lab 3

Task

We are going to be exploring the ethical idea of utilitarianism in this lab and specifically how changing the definition of your utility function can change the outcome of a decision. At a high level, the utility function is a measure of how much satisfaction an agent gets from a particular outcome.

You and four of your friends are deciding where to eat dinner in Ann Arbor. You have narrowed down the list to three restaurants: Cafe Zola, Frita Batido’s, and Tomokun. At this point, you each rank how much you want to go to each resaurant where 3 is your top choice and 1 is your bottom choice. These choices are stored in friends.csv.

friends.csv

Download the friends.csv file using the following command.

wget https://raw.githubusercontent.com/eecs298/eecs298.github.io/main/files/friends.csv

This file contains four columns: Person, CafeZola, Fritos, Tomokun. The Person column contains an ID for each friend and the other columns contains that friend’s ranking for the restaurant according to the scheme above.

lab3.py

Create a file named lab3.py in the same folder as the friends.csv file. Your task will be to choose a restaurant for the group using two different definitions of utility. First, read in the friends.csv file using the csv package and create a dict object where the key is the name ofeach restaurant and the value is a list of ranking values for that restaurant. Remember that csv reads in data as type str, so cast the ranking values to int when you read them in so that we can compute the utility later on.

Next, create a function named choose_restaurant that takes in the dict object you created above and a keyword argument utility_function which is a str that specifies which utility definition you should use in choosing a restaurant. The two definitions we will use are

Using this function, print out the top restaurant choice and utility value for each utility definition and the output you get when you pass in utility_function="other". Your output should be formatted as below.

Restaurant choice: XX, utility function: sum, value: XX
Restaurant choice: XX, utility function: top_choice, value: XX
Invalid utility function: other

Tips

Type casting

Python is a dynamically typed language, which means that the type of any variable will be determined during runtime. As such, you do not need to specify the type of a variable when you instantiate it. You still need to keep track of a variable’s type, though, as certain operations or functions will still require certain types. To cast a variable, use the built-in functions:

num = 123
num = str(num) # "123"
num = int(num) # 123
num = float(num) # 123.0
num = bool(num) # True

You can also the Python built-in functions type and isinstance to return the type of an object and determine if an object is of a certain type, respectively.

num = 123
type(num) # <class 'int'>
num = str(num)
isinstance(num, int) # False
isinstance(num, str) # True

String formating

When printing out values of variables in your Python program, it is helpful to use string formatting. If you put f before the "" in a string definition, then you can include variable names inside {} in the string definition to include their values. See the example below.

num = 298
string = "EECS"

formatted_string = f"Welcome to {string} {num}!"
print(formatted_string) # Welcome to EECS 298

List operations

Find the minimum or maximum value in a list with the min() and max() functions.

my_list = [4, 7, -2, 1, 3, 10]

min(my_list) # -2
max(my_list) # 10

Remove an item with a given value from a list with list.remove().

my_list = [10, 11, 12, 13]

my_list.remove(12)

my_list # [10, 11, 13]

Find the sum and length of a list using sum() and len().

my_list = [1, 3, 5, 7, 9]
sum(my_list) # 25
len(my_list) # 5

Conditionals

We can use conditionals statements similar to how you would in C++. Consider the following comparison examples.


if(condition1) {
    // Do stuff
}
else if(condition2 && condition3){
    // Do stuff
}
else{
    // Do stuff
}

is equivalent to:

if condition1:
    # Do stuff
elif condition2 and condition3:
    # Do stuff
else:
    # Do stuff

Iterables and loops

One python data type is the Iterable. This type is inherited by lists, tuples, and many others. A type that inherits from the Iterable type can be used as an iterator in for loops. For example:

my_list = [123, 456, 789]
for item in my_list:
  # Do stuff

The dict.keys() and dict.values() functions used in previous labs each return an object of the Iterable type. As such, they can be iterated through in for loops. You may have noticed that they are not actually lists, and thus cannot be used with list operators, unless explicitly cast to a list with list(dict.keys()).

You can use the range(n) function to create an Iterable over the natural numbers from 0 through n-1. With this you can construct for loops similar to those you would see in C++. For example:

for(int i = 0; i < n; i++){
  // Do stuff
}

is equivalent to:

for i in range(n):
  # Do stuff

You can also use two built-in functions enumerate(Iterable) and zip(Iterable1,Iterable2,...) to loop over an Iterable in a special way. See below.


myList1 = ["a","b","c"]
myList2 = [1,2,3]

for i, l1 in enumerate(myList1):
    print("Index:",i, "Value:",l1) 

for l1,l2 in zip(myList1, myList2):
    print("Value 1:", l1, "Value 2:", l2)

You would get the output

Index: 0 Value: a
Index: 1 Value: b
Index: 2 Value: c
Value 1: a Value2: 1
Value 1: b Value2: 2
Value 1: c Value2: 3

Another loop type while. This loop will continue as long as the conditional in the loop header evaluates to True.

while my_boolean: # Will continue as long as my_boolean is True
  # Do stuff

Other tools you can use while constructing loops are the continue and break keywords. In a for or while loop, the continue keyword will skip to the next pass through the loop, starting at the top of the loop. The break keyword will end the entire loop early.

for thing in other_thing:
  # Do stuff
  if certain_condition:
    continue # Skips to the next thing in other_thing
  else if other_condition:
    break # Ends the entire loop and moves on

  # Do other stuff -- this code will only run if neither continue or break are called.

List comprehension

You can create new lists using Iterables and for loops using something called list comprehension. See below for an example.


myList = [1,2,3]

myNewList = [element*2 for element in myList] # [2,4,6]

You can also include conditionals in the list comprehension to include only certain elements or apply operations to only some of the elements in the original Iterable. Note the difference in position of the if statement depending on whether an else statement is included or not.

myList = [1,2,3]

myNewList = [element if element == 2 else element**2 for element in myList] # [1,2,9]
myNewList2 = [element for element in myList if element == 2] # [2]

Dictionary comprehension

We can use dictionary comprehension the same way as list comprehension to create a dictionary.

keys = ["key1", "key2", "key3"]

dictionary = {k: 0 for k in keys} # intializes a dictionary where each key has value 0 to start