Data Mining Programming assignment #1

    • Python data structures and other essentials

1.1    Tuples

    1. Create the tuple named apps tuple with the string elements "Google", "Facebook", "Amazon", "Net ix", "AirBnB", "Instagram".

    2. Extract the second element by using its index and the last element by using its neg-ative index.

    3. Slice the elements from "Facebook" to "Net ix" inclusive on both ends.

    4. Try appending the element "Messenger" to the tuple. What do you observe? Why?



1.2    Lists

    1. De ne the list apps list with the same elements as the tuple above.

    2. Slice the rst three elements of the list with the shorthand syntax, i.e. not by typing explicitly all the indexes 0, 1, and 2; use the symbol : instead.

    3. Slice all the elements after index 3 inclusively with the shorthand syntax.

    4. Append the element ’Messenger’ to the end of the list and insert the element ’Youtube’ at index 1.

    5. Remove element ’Facebook’ and con rm with the function in whether ’Facebook’ has been removed from the list.

    6. Concatenate the last list with the list [’Linkedin’, ’Twitter’].









1
1.3    Dicts

    1. Create the dict named apps dict with the same values as in 1.1.1. and keys: app0, app1, etc.

    2. Access the element with key app1; then replace its value with ’Youtube’.

    3. Add a new key-value pair: app6-Messenger.

    4. What does the syntax: apps dict["app1"]="Messenger" do?

    5. What does the syntax: apps dict["App1"]="Facebook" do? Why?

    6. What does the syntax: del apps dict["App1"] do?



1.4    List comprehension

    1. Create the list named values with integers: 7, 12, 9, 18, 15.

    2. Create a list comprehension that takes the object values and returns the square of every value.


1.5    Functions and control    ow

    1. Generate a sample of 10 random integer numbers between 600 and 900 (hint: use Numpy’s random integers function).

    2. Create a function called credit score that reads the sample and returns the output ’Low’ when the input is [600, 699], ’Medium’ when in [700, 799] and ’High’ when in [800, 900].



    • Linear algebra in Numpy

2.1    Matrix calculations

1. Create matrices A and B as numpy arrays and calculate 2A, -3B and A+B
A =
21
13
B =
2 3
13

2
1

4
4
1
5

42
35


2
1





2

2. Create matrices A and B as numpy arrays and calculate A*B and the inverse of B

    • 3

=

2
6


=

2
3



1
5






A

4
4
0
5
B


1
1












3. Replace the values of the elements with abs(value) 3 with the value 3. Also, calculate the determinant of the original matrix A

    • 3
6
1
0
4
1


2
1
3
2


3
2
1
17
A =




7
6
0
0
0
2

4




5







2.2    Norms and eigenvalues

1. Calculate the maximum and Euclidean norm of the three vectors:
     T    T    T

x1 =  1    2    3    x2 =  2  0    1    2    x3 =  0  1    4    2    1

2. Calculate the l2 and l1 norms of the matrices:
A1= 4
3
A2
=
2 0
1   13




1
0
2




1
1
1
1
2


4

5








3. Calculate the eigenvalues of the matrix:

    • 3

= 4
2
2
3
A

10
1
6


10
2
95
    • Pandas manipulations

        1. Import le Gas prices.xls as a pandas DataFrame. Create a function that replaces NaN with the interpolation between the adjacent values (i.e. the prior and the subse-quent to the NaN).

        2. Find the index, the date and the value of the min and max gas price.

        3. Calculate the following descriptive statistics: mean, median, quantile, skewness and kurtosis.




3

    4. Create two new columns with the Month and Year. Pivot the table by Month(index) and Year(columns). Calculate and plot the monthly average gas price. This should give you one line per year.

    5. Create a new column with the season of the year, set it as a secondary index and calculate the average of every season (regardless of the year).


    • Pivot, aggregate and plot timeseries

        1. Import le TREB data.xls as a pandas DataFrame. Create the necessary columns to pivot and generate the table below with the value of Sales (hint: use the split() function to split the MonthYear column to Month and Year).

        2. In one gure, plot 5 lines (one line per year) with the monthly sales versus the month of the year.

        3. Calculate the total Sales per year and create a bar chart. Do this in two di erent ways: a) group by year the table pre-pivoting, b) summing up the ve columns post-pivoting


Sales
Year

Month
2013
2014  2015  2016  2017



January


.


.


.


December
































4