ENGG2112 Coding Assignment
Due on 23 April 2023, 11.59pm
10 APRIL 2023
Instructions
This is an individual assignment and the submitted work must be your orig
inal work. You are allowed to discuss the method of solution with others,
however the submitted code must be entirely written by you.
Submit your work as a Python notebook in the template provided.
Submissions must be made through Canvas only, and not by e-mail. The
deadline will be strictly enforced: 11:59pm on 23 April 2023. (Students with
disability adjustments will be contacted separately.)
Please plan your time according to your own ability and schedule, seek help
from the teaching team and peers in a timely fashion, and try not to ask for
deadline extensions.
Download the fifile House_Rent_Dataset.csv from the Canvas website. Use
the notebook template ENGG 2112 coding assignment 2023 (1).ipynb to an
swer the following questions.
12
Description of the Dataset
In this Dataset, we have information on almost 4700+ Houses/Apartments/Flats
Available for Rent with different parameters like BHK, Rent, Size, No. of Floors,
Area Type, Area Locality, City, Furnishing Status, Type of Tenant Preferred, No. of
Bathrooms, Point of Contact.
Dataset Glossary
BHK Number of Bedrooms, Hall, Kitchen
Rent Weekly Rent of the Property
Size Size of the Property in Square Feet
Floor Floor location of property and total number of flfloors in building (e.g. Ground
out of 2, 3 out of 5, etc.)
Area Type Size of the property calculated on either Super Area, Carpet Area or
Build Area.
Area Locality Locality of the Property
City City where the Property is located
Furnishing Status Furnished, semi-furnished or unfurnished
Tenant Preferred Type of tenant preferred by the owner or agent
Bathroom Number of bathrooms
Point of Contact Person to contact for more information
Problem 1
(This problem is worth 2 marks in each part, 6 marks in total.)
1. Find the minimum, maximum and average rent in the entire dataset. Assign
these values to the variables rent_min, rent_max and rent_avg respectively.
2. Find the subset of data records that satisfifies the following conditions:
? The posted date is in June 2022 (i.e. 1st to 30th June 2022).
? Information on the property should be obtained from the agent.
? The size of the property is at least 1,000 square feet.
Create the dataframe df_q2 to hold the data, and determine the number of
eligible records/samples. Put this value in the variable num_rows.3
3. In the fifirst cell, plot a histogram of property sizes. In the second cell, plot a
scatter plot of property size versus date of posting. Ensure that the date is in
ascending (i.e. chronological) order.
Problem 2
1. (3 marks) Use the columns “BHK", “Size", “Area Type" and “Bathroom" to
build a linear regression model to predict the rent of a property. Convert all
the categorical data into binary variables using one-hot encoding. Use 75% of
the data for training, with the random state set to 2112. Find the coeffificient of
determination R2 and the mean squared error, and store these values in the
variables R2 and mse respectively.
2. (6 marks) Use the columns “BHK", “Size", “Floor", “Area Type" and “Bath
room" to build the following three classififiers to predict the furnished status
of the property:
a) Logistic regression with max_iter = 1000.
b) Multi-layer perceptron with one hidden layer of 100 neurons, maximum
number of iterations = 500, and random state = 2112.
c) Gaussian Na?ve Bayes
Process the data as in Problem 2.1 above. In addition, for the “Floor" column,
extract the information to two new columns “Floor_new" and “Total_flfloor”,
containing the flfloor location of the property and the total number of flfloors
in the building, respectively. Transform the “Floor” information as follows:
Ground → 0, Upper Basement → ?1, Lower Basement → ?2. Insert the two
new columns into the dataframe and delete the original column “Floor".
Compare the performance of the three classififiers on the test data. The evalu
ation metrics are f1 score and accuracy, both stored in the variable result.
Problem 3
(5 marks) Use the columns “BHK”, “Size” and “Bathroom” in a K-nearest neigh
bours (KNN) predictor of rent and furnished status. The model needs to be built
from scratch, i.e. without using any pre-packaged function that implements KNN
in existing libraries. The data should fifirst be pre-processed using min-max normal
ization, i.e. replace each feature xi , with minimum and maximum values across the
dataset of xi,min and xi,max, with the normalized feature.
Test your function knn using the two sample data records provided in the last two
cells of the Python answer notebook.