ECBS5294 · Session 1

Working with Data

ECBS5294 · DS2 · MSBA 2026–2027

Before Block 1: what this course is, what is new, how it is graded

Welcome — the course, the people, and how it is graded
ECBS5294 · Session 1

What this course is about

No error is not the same as correct.

  • In DS1 almost everything that went wrong crashed: a missing file, the wrong Python, a stale kernel.
  • In DS2 almost nothing crashes. The query runs, the number comes out, and the number is wrong: a join tripled it, a missing value vanished, the world was counted six times.
  • The skill: look for evidence that a number is wrong before anyone complains, then build the check into the code so it fails out loud next time.
Welcome — the course, the people, and how it is graded
ECBS5294 · Session 1

Three Mondays, three questions

Session Date Question Homework
1 Mon 5 Oct What is this data? — grain, keys, types, missing values; one table in SQL HW1 due Sat 10 Oct, 23:59
2 Mon 12 Oct What happens when tables meet? — joins; APIs and JSON into tables HW2 due Sat 17 Oct, 23:59
3 Mon 19 Oct Can I trust the answer? — bronze, silver, gold; checks that can fail HW3 due Sat 31 Oct, 23:59

Final exam: closed-book (one handwritten A4 sheet allowed), in person, about 70 minutes, in a separate session after HW3. The date is on Moodle. A mock exam with a key comes first.

Under all of it, DS1's loop: observe → locate → hypothesize → inspect → change → verify → explain.

Welcome — the course, the people, and how it is graded
ECBS5294 · Session 1

Real data, every time

Data What it is Where you meet it
Online Retail II 525,461 invoice lines of a UK gift retailer, 2009–2010 Block 1 lecture, Lab 2, HW1
Olist a Brazilian marketplace: 2,653 of its 99,441 orders, selected to keep every problem the labs need Labs 3, 4, 6, HW2
Our World in Data CO₂ and greenhouse-gas emissions, every country, every year Lab 1, HW3
Eurostat EU waste and emissions, exactly as Eurostat publishes them Block 5, Lab 5, HW3
ECB via Frankfurter · World Bank Two real APIs, cached so no lab needs the network Block 4, Lab 4, HW2, HW3

Nothing is invented and nothing is edited. Every problem you find in a lab is something the real file does.

Welcome — the course, the people, and how it is graded
ECBS5294 · Session 1

What is new, and what is not

The same as DS1: the rhythm (45 lecture · 5 stretch · 45 lab), the diagnosis note, explain it to a neighbour at minute 33, the Moodle checkpoint, labs carry no points, Git Bash on Windows.

New in DS2:

  • In the lecture you predict, then watch. every query is on a slide; first you write down what you expect, and one reason why, then the next slide shows what it printed. You type in the lab. Each lab's clone starts in the stretch before it.
  • The lab has two halves: find why a number is wrong, then write the queries the report is missing. At least one of them from an empty cell.
  • SQL on DuckDB is the language. Every lab's uv sync installs it.
  • Labs 3 and 5 open with 20 minutes of no AI.
Welcome — the course, the people, and how it is graded
ECBS5294 · Session 1

How you are graded

Component Weight
Homework 1 — One table, the weekly numbers 10%
Homework 2 — Many tables and an API: the board-deck numbers 15%
Homework 3 — Handover: the emissions brief 15%
Knowledge checks (start of Sessions 2 and 3, closed-book) 10%
Final exam (closed-book, in person, separate session) 50%

The same as DS1. The exam asks you to write SQL on paper, alone: four queries, 40 of its 100 points. The labs are the practice for it.

Welcome — the course, the people, and how it is graded
ECBS5294 · Session 1

AI: it writes SQL that runs. Running is not right.

  • Yes: explain this error; explain this clause; suggest what to check; improve my writing.
  • No: submit a query or a note you cannot defend in person; describe work you did not do.
  • An assistant given your schema writes a join that returns rows. It will not tell you the rows are tripled. The counts are yours to run.
  • The habit: paste the schema and the grain, the number you expected and the number you got — and ask for the checks that would tell them apart, not for the fix.
Welcome — the course, the people, and how it is graded
ECBS5294 · Session 1

Who is here

  • Eduardo Ariño de la Rubia — RubiaE@ceu.edu · office hours by appointment.
  • Your teaching assistants — introducing themselves now. During labs: questions, not answers.
  • Every one of us has shipped a wrong number that ran without an error. Thirty seconds each.
Welcome — the course, the people, and how it is graded
ECBS5294 · Session 1

The rules that bite this week

  • Late homework: one day late at −10%; nothing after Sunday 23:59, because the key posts on Monday morning, before the knowledge check.
  • Extensions: email me before the deadline, with a reason. Documented illness or emergencies are always accommodated.
  • Accommodations: the skill is diagnosis, not speech. The explanation to a neighbour can be written — arrange it with me this week.
  • Every other rule is the same as DS1's, and the syllabus has them all. Moodle is authoritative for every date.
Welcome — the course, the people, and how it is graded