Map, Filter and Lambda in Python Lab

Objectives

  • Apply and combine the skills covered for map and filter functions
  • Learn how to write and use lambda functions for transforming data
  • Modify given data using map and lambda functions as an alternative to writing for loops
  • Filter given data using filter function to only include the data that meets a given criteria

Introduction

In this lab, we'll put our new knowledge about map and filter to the test. We'll also introduce lambda functions as a convenient tool for transforming data on the fly. As a test case, we'll be working with Yelp data again. Let's get started!

Lambda functions

Recall that map applies a given function to every element of an iterable. Previously, you've seen map used with a variety of built-in Python functions. As you begin to work with more complicated data, you may need to use a custom function that performs a unique task for which there is no built-in Python function. This is exactly what lambda functions are used for.

Say you wanted to add 5 to every element in the list of numbers shown below:

# List of numbers
numbers = [1, 3, 8, 9, 11, 20]
numbers
[1, 3, 8, 9, 11, 20]

Unfortunately, you can't use the addition operator as this results in a TypeError:

numbers + 5
---------------------------------------------------------------------------

TypeError                                 Traceback (most recent call last)

<ipython-input-2-be56e3573393> in <module>()
----> 1 numbers + 5


TypeError: can only concatenate list (not "int") to list

If there were a built-in Python function to add 5, you might just use that function with map to apply it to all of the numbers in the list. But, sadly, no such function exists. The good news is that lambda can be used to define a custom function that adds 5! The syntax for defining a lambda function that adds 5 is shown below:

lambda x: x + 5

As you might have guessed, x here is a variable and the lambda function simply adds 5 to it. Now that you understand how to write lambda functions, use map to apply the lambda function above to add 5 to every number in the numbers list.

list(map(lambda x: x + 5, numbers))
[6, 8, 13, 14, 16, 25]

Note that you don't always have to use x as the variable. You can define the variable with any name you want as long as the syntax is correct!

The cool thing is that lambda functions are customizable. They are not just limited to numeric applications. You'll see how lambda functions can be used to transform text data below with the Yelp data set.

Yelp data

Now that you've been introduced to lambda, you can practice using it with map and filter to handle some real-world data. Let's start with the Yelp restaurants data set. The code below uses lambda to create a dictionary consisting of 4 keys: name, price, is_closed, and review_count. The map function is then used to apply the lambda function to every restaurant in the data set.

from restaurants import yelp_restaurants # in this line we are simply importing our data we gathered from Yelp.
restaurants = list(map(lambda restaurant: dict(name=restaurant['name'], 
                                           price=restaurant['price'], 
                                           is_closed=restaurant['is_closed'],
                                           review_count=restaurant['review_count'],
                                          ), yelp_restaurants))

We now have a list of restaurants from the Yelp Api. Let's take a look at the list.

restaurants

Using map

As you can see, it's a little tricky to see the names of all of the restaurants due to amount of data. Let's create a new list names to contain only the names of all the restaurants from the list above. Use the map and lambda functions, along with your understanding of a dictionary's structure to do so.

names = None
names
# ['Fork & Fig',
#  'Salt And Board',
#  'Frontier Restaurant',
#  'Nexus Brewery',
#  "Devon's Pop Smoke",
#  'Cocina Azul',
#  'Philly Steaks',
#  'Stripes Biscuit']

This worked well. Now let's get a sense of how many reviews were written for each of these restaurants. Just like above, create a new list review_counts to only contain the values of review_count for each restaurant.

review_counts = None
review_counts # [610, 11, 1373, 680, 54, 647, 25, 20]

Let's say we want to get a sense of total number of reviews in the whole dataset. We can add up the elements in review_counts list, and assign the result to a variable named total_reviews.

total_reviews = None
total_reviews # 3420

It's a little tricky to work with the price in the format of dollars signs i.e. $ and $$ based on how expensive the restaurant is.

So write a function called format_restaurants that changes each restaurant to have the attribute 'price' point to the number of dollar signs (i.e. 1 for $ and 2 for $$). We'll get you started with the function, format_restaurant.

def format_restaurant(restaurant):
    if type(restaurant['price']) == str:
        restaurant['price'] = len(restaurant['price'])
    return restaurant
format_restaurant(restaurants[0]) # {'name': 'Fork & Fig', 'price': 2, 'is_closed': False, 'review_count': 610}

Now write another function called map_format_restaurants using map, that uses above function and returns a list of restaurants with each of them formatted with price pointing to the respective number.

def map_format_restaurants(restaurants):
    pass
map_format_restaurants(restaurants)

#[{'name': 'Fork & Fig', 'price': 2, 'is_closed': False, 'review_count': 610},
# {'name': 'Salt And Board',
#  'price': 2,
#  'is_closed': False,
#  'review_count': 11},
# {'name': 'Frontier Restaurant',
#  'price': 1,
#  'is_closed': False,
#  'review_count': 1373},
# {'name': 'Nexus Brewery',
#  'price': 2,
#  'is_closed': False,
#  'review_count': 680},
# {'name': "Devon's Pop Smoke",
#  'price': 2,
#  'is_closed': False,
#  'review_count': 54},
# {'name': 'Cocina Azul', 'price': 2, 'is_closed': True, 'review_count': 647},
# {'name': 'Philly Steaks', 'price': 2, 'is_closed': False, 'review_count': 25},
# {'name': 'Stripes Biscuit',
#  'price': 2,
#  'is_closed': True,
#  'review_count': 20}]

Filter

Now let's use filter to search for restaurants based on specific criteria.

Write a function called open_restaurants using filter and lambda that takes in a list of restaurants and only returns those that are open. You can use the distionary key is_closed to make a decision in your code.

def open_restaurants(restaurants):
    pass
open_restaurants(restaurants)

#[{'name': 'Fork & Fig', 'price': 2, 'is_closed': False, 'review_count': 610},
# {'name': 'Salt And Board',
#  'price': 2,
#  'is_closed': False,
#  'review_count': 11},
# {'name': 'Frontier Restaurant',
#  'price': 1,
#  'is_closed': False,
#  'review_count': 1373},
# {'name': 'Nexus Brewery',
#  'price': 2,
#  'is_closed': False,
#  'review_count': 680},
# {'name': "Devon's Pop Smoke",
#  'price': 2,
#  'is_closed': False,
#  'review_count': 54},
# {'name': 'Philly Steaks', 'price': 2, 'is_closed': False, 'review_count': 25}]

Let's say we now want to look at restaurants that are comparatively cheaper i.e. $ or 1 as price.

Write a function called cheap_restaurants using filter, that returns the restaurants that have a price of 1, or '$'.

def cheapest_restaurants(restaurants):
    pass
cheapest_restaurants(restaurants)

# [{'name': 'Frontier Restaurant',
#  'price': 1,
#  'is_closed': False,
#  'review_count': 1373}]

So we have only one restaurant in the data that meets the given criteria. Next, we shall write a function that filters out only those restaurants that 100 reviews or more, since we want to make sure there is some solid data points backing the reviews -- we are burgeoning data scientists after all!

def sufficiently_reviewed_restaurants(restaurants):
    pass
sufficiently_reviewed_restaurants(restaurants)

# [{'name': 'Fork & Fig', 'price': 2, 'is_closed': False, 'review_count': 610},
# {'name': 'Frontier Restaurant',
#  'price': 1,
#  'is_closed': False,
#  'review_count': 1373},
# {'name': 'Nexus Brewery',
#  'price': 2,
#  'is_closed': False,
#  'review_count': 680},
# {'name': 'Cocina Azul', 'price': 2, 'is_closed': True, 'review_count': 647}]

Summary

Neat! In this lab, we successfully proved our prowess when it comes to iterating over each element of a list with both map and filter! We also learned about lambda functions and how to use them. We used map to format our data into ways that better help us answer questions and extrapolate insights. We used filter to return subsets of our data like our restaurants that were only one $ or our restaurants that had 100 or more reviews.