Correction of 1st exam 2023/24

NES—Programmiertechniken (Introduction to scientific programming)

19.01.2023

Please work through the following tasks. You have 75 minutes to complete the exam.
Make sure you executed all relevant code cells and save the notebook before the end of the exam.

This is a correction of the exam questions. Like always in programming, there are many ways to achieve similar results. My suggestions here follow our coding exercises discussed in class.

If you want to solve this exam as an exercise, download the uncorrected notebook.

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

Question 1: Multiple-choice (5 points)

Careful: wrongly ticked answers will give negative points!

  1. Which of the following statements are true? (2 points)
  • The python interpreter is a program that can read and execute code written in the python programming language.
  • There are several python interpreters with several versions. ipython is one of these interpreters.
  • The python open-source ecosystem provides a large variety of packages, which are ready-made bundles of code that you can install and use.
  • Everyone can write and distribute their own python package to be used by others.
  • Using conda you can create an integrated development environment to manage the installation of python and python packages you might need.
  1. Which of the following statements are true about this line of code: (2 points)
    conda create -n scipro2023 python=3.11 ipython jupyterlab numpy pandas scipy matplotlib
  • This is not python code.
  • This creates a conda environment with the name scipro2023.
  • This triggers some installations. For example, python version 3.11 will be installed.
  • After running this line, you can not install any other versions of python on the same laptop anymore.
  • numpy, pandas, scipy, and matplotlib are python packages.

Question 2: datetime calculations (5 points)

Your working directory contains a python module named solar_exam. You know most of the module already from our lecture assignments. I just added one more function to it.

Solve the following tasks:

  1. Create a DataFrame df with a DatetimeIndex ranging from 2024-01-01 00:00 to 2024-01-31 23:00 in hourly sampling. (1 point)
datetime = pd.date_range('2024-01-01 00:00', '2024-01-31 23:00', freq='H')
df = pd.DataFrame(index=datetime)
  1. Use df and the function compute_solar_dataframe() contained in solar_exam to fill your data frame with data columns. (1 point)
import solar_exam as solar
df = solar.compute_solar_dataframe(df)
print(df)
                     hourangle  declination  elevangle     azimuth  radiation
2024-01-01 00:00:00     -180.0   -22.930544 -65.518144         NaN        0.0
2024-01-01 01:00:00     -165.0   -22.930544 -62.729858  211.348308        0.0
2024-01-01 02:00:00     -150.0   -22.930544 -55.750441  234.906273        0.0
2024-01-01 03:00:00     -135.0   -22.930544 -46.681349  251.665818        0.0
2024-01-01 04:00:00     -120.0   -22.930544 -36.760543  264.601524        0.0
...                        ...          ...        ...         ...        ...
2024-01-31 19:00:00      105.0   -17.516495 -22.868690   88.634290        0.0
2024-01-31 20:00:00      120.0   -17.516495 -32.974356  100.117762        0.0
2024-01-31 21:00:00      135.0   -17.516495 -42.681114  113.470138        0.0
2024-01-31 22:00:00      150.0   -17.516495 -51.303953  130.299509        0.0
2024-01-31 23:00:00      165.0   -17.516495 -57.665833  152.518291        0.0

[744 rows x 5 columns]
/home/flo/miniconda3/envs/scipro2023/lib/python3.11/site-packages/pandas/core/arraylike.py:396: RuntimeWarning: invalid value encountered in arccos
  result = getattr(ufunc, method)(*inputs, **kwargs)
  1. The resulting DataFrame likely contains some NaN. Which columns contain how many NaN’s? (1 point)
df.isna().sum()
hourangle       0
declination     0
elevangle       0
azimuth        29
radiation       0
dtype: int64
  1. Resample the two columns declination and elevangle to daily sampling by computing each day’s average values.
    Create a quick working plot of these average values with as few lines of code as possible. The resulting figure looks like that: (2 points)

dfr = df[['declination', 'elevangle']].resample('D').mean()
dfr.plot()
<Axes: >

Question 3: Spreadsheet data (8.5 points)

Your working directory contains a spreadsheet data set about the hydropower consumption of different countries over several decades.
Solve the following tasks:

  1. Load the data contained in the spreadsheet dataset.csv. (1 point)
hc = pd.read_csv('dataset.csv')
print(hc)
          Country Code  Year  Electricity
0     Afghanistan  AFG  2000         0.31
1     Afghanistan  AFG  2001         0.50
2     Afghanistan  AFG  2002         0.56
3     Afghanistan  AFG  2003         0.63
4     Afghanistan  AFG  2004         0.56
...           ...  ...   ...          ...
7569     Zimbabwe  ZWE  2016         2.95
7570     Zimbabwe  ZWE  2017         3.93
7571     Zimbabwe  ZWE  2018         5.00
7572     Zimbabwe  ZWE  2019         7.26
7573     Zimbabwe  ZWE  2020         7.26

[7574 rows x 4 columns]
  1. How many unique countries does the data set contain? (1 point)
hc['Country'].unique().shape[0]
224
  1. How many records are contained in the data set where the electricity is greater or equal to 100 but below 250 TWh? (1 point)
np.sum((hc['Electricity'] >= 100) & (hc['Electricity'] < 250))
215
  1. Sort the countries according to most hydropower consumed. Use the DataFrame methods .groupby() and .sort_values() for this task. (1.5 points)
hcm = hc.groupby(['Country']).max()
hcm.sort_values('Electricity', ascending=False)
Code Year Electricity
Country
Asia NaN 2020 1958.06
China CHN 2021 1322.01
Europe NaN 2021 775.34
Latin America and Caribbean NaN 2021 765.88
North America NaN 2021 746.22
... ... ... ...
Solomon Islands SLB 2020 0.00
Cyprus CYP 2021 0.00
South Sudan SSD 2020 0.00
Cook Islands COK 2020 0.00
Saint Lucia LCA 2020 0.00

224 rows × 3 columns

  1. Plot the electricity curves of the two countries with most hydropower consumption. Make your figure look as similar as possible to the following one.
    If you don’t know which countries did consume most hydropower, use Canada and India. (2 points)

country1 = 'Asia'
country2 = 'China'

fig, ax = plt.subplots()
ax.plot(hc.loc[hc['Country'] == country1, 'Year'], hc.loc[hc['Country'] == country1, 'Electricity'], label=country1)
ax.plot(hc.loc[hc['Country'] == country2, 'Year'], hc.loc[hc['Country'] == country2, 'Electricity'], label=country2)
ax.legend()
ax.grid(linestyle='dotted')
ax.set_ylabel('Electricity (TWh)')
ax.set_title('Hydropower consumption')
plt.show()

  1. Not all countries are associated with a country code. Iterate through all unique countries to find out which countries are not associated with a code. Your loop should
      1. print a list of countries whose country code is NaN and
      1. count and display how many countries those are.

The DataFrame method .isin() and the numpy NaN np.nan will be helpful.
(2 points)

counter = 0
for country in hc['Country'].unique():
    if hc.loc[hc['Country'] == country, 'Code'].isin([np.nan]).any():
        print(country)
        counter = counter + 1

print(counter)
Africa
Asia
Eastern Africa
Europe
European Union (27)
Latin America and Caribbean
Low-income countries
Lower-middle-income countries
Middle Africa
Middle East
North America
Oceania
South America
Western Africa
14

Question 4: Functions (5 points)

The following code cell contains a function signature with its documentation. Code the according function body.
Tip: You can use the numpy function np.floor() to help you separate the floating point number into its whole number and the decimal number. Consult the interactive documentation of np.floor() if you need more explanation or want to look at examples.

def pretty_print_hours(hours):
    """Convert a floating point representation of hours into a more human readable string format of 'hours:minutes'.
    The function does not allow negative inputs!

    Parameters
    ----------
    hours: float
        Hours to convert (No negative numbers allowed!)

    Returns
    -------
    out: string
        Hours in the format 'hours:minutes'

    Examples
    --------
    >>> pretty_print_hours(1.5)
    '1:30'
    >>> pretty_print_hours(-2.3)
    Traceback (most recent call last):
     ...
    ValueError: 'hours' cannot be negative!
    """
    if hours < 0:
        raise ValueError("'hours' cannot be negative!")

    hours_int = int(np.floor(hours))
    min_decimal = hours - hours_int
    min_int = int(min_decimal * 60)

    out = f'{hours_int}:{min_int}'
    
    return out
pretty_print_hours(1.5)
'1:30'
pretty_print_hours(-2.3)
ValueError: 'hours' cannot be negative!

Question 5: Debugging (3 points)

Let’s first define three variables:

count = np.arange(10, 20)
v1 = np.full(count.shape, False)
v2 = np.full(count.shape, False)

Now, read through the next code block that uses logical indexing to change some values of v1:

v1[(count > 13) & (count < 17)] = True

The following code block tries to re-write the logical indexing into a loop.
Unfortunately, there are three mistakes in the next code block.
Fix the mistakes, so that v1 and v2 are equal.

# Code cell with mistakes:
for ct in enumerate(count):
    if ct > 13 and ct < 17:
    v2[ct] = True
# Corrected code cell:
for i, ct in enumerate(count):
    if ct > 13 and ct < 17:
        v2[i] = True

Test whether v1 and v2 are equal:

np.array_equal(v1, v2)
True
# Corrected code cell (alternative strategy applied by several students):
for ct in range(len(count)):
    if count[ct] > 13 and count[ct] < 17:
        v2[ct] = True