Description
1. (25 points) Numpy Exercise: Create a numpy matrix with random numbers (use the
numpy.random.randn function) with 5 rows and 5 columns. Compute the following functions for
this matrix.
a. Rescale each element such that it lies in the range between (-1,1). That is the values in
the i-th column needs to be rescaled with respect to the min and max values in the i-th
column.
b. Standardize the elements of the matrix. That is, for each element at position (i,j), divide
the element by the mean of the i-th column and divide by the standard deviation of the
i-th column.
Display the original and the changed matrices
2. Pandas Exercise (25 points) Load the CSV file named “311-service-requests.csv” as a pandas
dataframe and perform the following
a. We want to find compare complaints for different cases for each “Borough” (one of the
feature-names in the csv). Specifically, Draw a bar chart that compares for each
“Borough” the total number of complaints where the “Complaint Type” includes the
phrase “Noise”, and the total number of complaints where the “Complaint Type”
includes the phrase “Parking”. (Hints: Look at function value_counts() for this problem)
3. Pandas Exercise (50 points): Load the CSV file named “weather-2012.csv”
a. What month of the year has the ideal temperature? To answer this question, draw the
average temperature (“Temp (C)” column) for each month in the year as a line graph
(25 points)
b. Distribution Fitting: What is the natural distribution for the feature “Wind Spd (km/h)”?
To answer this, fit three distributions, normal, logormal and beta for this feature. Show
the histogram for this feature assuming 10 bins and all the fitted distributions on the
histogram. (25 points)
Please write your code using ipython and submit your ipynb file to dropbox on ecourseware.
