This course was about programming “the right way”

Program in a way that is reproducible, reusable and collaborative

• Produce code that works not just on your machine, but in settings like a research team or industry data job.
• Develop a mindset
• Similar to math, statistics, and Econ 101, develop a toolkit that will serve you in your studies and in your career 🧰

The course

The course

The course

The course

The course

Iteration # 1

• Version 1.0 of this course
• We hope you found this course stimulating, valuable, and directly applicable to your studies/career
• For us, it has been an intense yet rewarding, engaging and fun experience

General remarks

• Digital examination, 60mn, during class on Friday, May 22nd.
• BYOD, make sure you bring your laptop charging cable 🔋
• The lock-down browser must be installed before the exam. For any questions, refer to StudentWeb.

Structure

• True/False questions: one statement, decide if the statement is true or false.
• Multiple correct: one statement, 3-4 possible answers, decide which answers are correct.
• Only one correct: one statement, 3-4 possible answers, only one correct.
• Essay questions on the group project. A maximal word count is given (answers with higher word counts will have 10% points removed)

The type of the question is always indicated.

What to expect and how to prepare

What to expect

• Questions on the course and exercises, focus on understanding and on the intuition you developed through practice (testing, coding, etc. )
• Slides noted with 👓 in lectures 9 and 10 are not exam relevant.
• You are not asked to write down code, but you must be able to interpret code that we give you.

How to best prepare

• If you solved the exercises, attended the lecture, and understood your group project, you are ready.
• Make sure you understand the key concepts.
• Since this is the first time we give the course, there is no mock exam available.
• BUT: we will be generous in the grading to account for this.

Example: True-False question

After running git commit, your changes are immediately visible to your teammates on GitHub.

True or false

🔳 True

🔳 False

Example: Multiple correct question

class BankAccount:
    def __init__(self, owner_name):
        self.owner_name = owner_name
        self.balance = 0

    def withdraw(self, amount):
        if amount <= self.balance:
            self.balance = self.balance - amount
            print(f"New balance: ${self.balance}")
        else:
            print("Withdrawal denied.")

    def deposit(self, amount):
        self.balance = self.balance + amount

# Example
my_account = BankAccount(owner_name="Franziska")
my_account.deposit(30)
my_account.withdraw(50)
my_account.deposit(100)
print(my_account.balance)

Example: Only one correct question

It does not matter whether I do a LEFT JOIN or an INNER JOIN from my fact table to a dimension table in a star schema where relational integrity holds.

Only one correct.

1. The statement is incorrect: a LEFT JOIN always returns more rows than an INNER JOIN.
2. The statement is correct: relational integrity guarantees every row in the fact table has a matching row in the dimension table, so neither join drops rows.
3. The statement is incorrect: an INNER JOIN could silently drop rows from the fact table if a foreign key has no match in the dimension table.

Example: Essay question

In the group project, you were asked to use pull requests when working on github. Explain (1) how pull requests work, and (2) what are the advantages of using them, and (3) an example from your group project. Answer in max 70 words.

Feedback

• Feel free to fill out the course evaluation form. Constructive feedback is welcome.
• Approach us if you would like to share feedback, and/or ideas for the next iteration.

Exploring Inequality

Data: World Bank’s World Development Indicators (WDI)

We get the data from the World Bank “Bulk Download”

1. On the Worldbank Website, click Download > Bulk Download > CSV.
2. The WDI bulk download contains all indicators in one csv archive. Unzip the archive into week_11/data/.

The database

• We provide the ready-made database wdi.db, you can use it directly
• We also provide build_wdi.py, the script used to create it, in case you are curious how the ETL process was done

Design

The data is more “messy” than the CEPII dataset or the FRED dataset.

The entities:

• A country has an ISO code, a name, a region, and an income group
• An indicator has a WDI code, a name, a topic, and a unit
• An observation records one value for one country, one indicator, and one year

Classic 🌟 schema!

The model

The Class

• A class that connects to the database and produces visualisations of inequality across countries
• We show three versions that differ in how and when they retrieve data from the database
• The core logic (querying, plotting) stays the same — only the connection strategy changes
• All examples are in db_connection_examples.py