5. Download the Pittsburgh city fire incident data from the class web page. Using Jupyter notebook, write a Python program that performs the following steps.

(a) Use pandas to load the CSV file into a DataFrame object named data. After loading the data successfully:

i. Type data.head() to return the first 5 records of the dataset.

Solution:

In [ ]:

import pandas as pd data = ... data.head() 

ii. Determine the number of rows (data records) and columns (attributes) of the dataset using the data.shape command.

Solution:

In [ ]:

iii. Count the frequency of each type of fire incident by applying the value_counts() function to the type description column (see the pandas example given in lecture 2).

Solution:

(b) Using pymongo, write the Python code for storing the data into MongoDB. Specifically, create a collection named fire, where each document corresponds to a fire incident record. You must first convert each record into a dictionary object before inserting it into MongoDB (see the example given in Exercise 3). Each fire incident record should be stored as a set of key-value pairs in BSON format on MongoDB, where the keys should correspond to the attribute names of the original CSV data file.

Solution:

(c) Using pymongo, execute a MongoDB query that identifies the top-5 most frequent type of fire incidents. Your result should be consistent with the answer for Question 5(a)(iii). Hint: you will need to use the following MongoDB operators for the query: group, sort, and $limit. See the example given in https://www.mkyong.com/mongodb/mongodb-aggregate-and-group-example/ on how to write queries with aggregate functions. Note that aggregate functions are implemented in pymongo as follows:

db..aggregate( ... ) 

Solution: