Run your first lesson

Before you start

• Level: 1 (beginner)
• Time: about 1 hour
• Builds on: nothing. This is the first lesson of the course. The phase overview lists what comes next.
• You need: a computer (Windows, macOS or Linux) and an internet connection once, to download Python and this course. After that everything runs offline.

The problem

Tiffinbell is a fictional food delivery app with a kitchen in Pune. Today is your first day as its analyst. Ritu, who runs the kitchen, has a call with her vegetable supplier at 9 am. Before the call she wants two numbers from yesterday: how many orders the kitchen made, and how much money they brought in. The orders are in one small file, data/orders.csv. You will get Python working, run your first test, watch it fail, fix the code and watch it pass.

In plain words

Think of a lesson as a homework sheet that marks itself. The README is the teacher explaining the topic. The exercise file is your answer sheet with blanks to fill in. The test is the answer key, kept by a strict marker who works out the right answer on their own every time and tells you exactly which line is wrong.

A few words you will meet:

• Python: a free program that runs instructions written in the Python language. Analysts use it to read files and do sums that would take hours by hand.
• Terminal: a window where you type commands instead of clicking. On Windows it is called PowerShell or Terminal.
• Folder and path: a folder holds files. A path is the address of a folder, such as Desktop/analytics-from-scratch.
• Test: a small program that runs your code and checks its answers. It says OK when everything is right and FAILED when something is not.
• CSV: a "comma separated values" file. It is a spreadsheet saved as plain text: one line per row, commas between the cells.

What you will build from scratch

• A working Python 3.9 or later on your computer, checked with one command.
• The habit of opening a terminal and moving into a lesson folder with cd.
• count_orders: counts the orders in a list.
• total_amount: adds up the amount column with a loop.
• The skill of reading a failing test message and knowing what to fix.

The data in data/orders.csv is synthetic (made up by make_data.py with a fixed random seed, so everyone gets the same 38 rows). Amounts are in Indian rupees.

The idea

Every lesson in this course is built the same way.

The test does not copy the answer from the solution. It works the answer out again in plain Python, from the data. It also builds a second dataset with different numbers and runs your code on that too. So typing return 38 would pass the first check and fail the second. The only way to pass is code that really works.

This is the loop you will repeat in every lesson:

flowchart TD
    A["Read README.md"] --> B["Fill in the TODO lines<br/>in the exercise file"]
    B --> C["Run python3 test_lesson.py"]
    C --> D{"OK or FAILED?"}
    D -- FAILED --> E["Read the last line<br/>of each failure"]
    E --> B
    D -- OK --> F["Answer Check yourself<br/>out loud"]
    F --> G["Go to Next"]

Walk through

Step 1. Open a terminal

• Windows: press the Start button, type PowerShell and open Windows PowerShell (or Terminal, if you have it).
• macOS: press Command and Space together, type Terminal and press Return.
• Linux: open Terminal from your applications menu. On Ubuntu the shortcut is Ctrl, Alt and T together.

You see a line ending in $, % or >. That is the prompt: the terminal is waiting for you to type a command and press Enter.

Step 2. Check for Python 3.9 or later

Type one of these and press Enter:

python3 --version        (macOS and Linux)
py --version             (Windows)

You want an answer such as Python 3.9.6 or Python 3.12.4: the first number 3, the second number 9 or higher. On Windows, if py is not found, try python --version.

If you see "command not found", "not recognized", or a version below 3.9, install Python:

• Windows: download the latest Python 3 from the official page, https://www.python.org/downloads/ , and run the installer. If it shows a box called "Add python.exe to PATH", tick it. Then close PowerShell, open a new one, and run py --version again.
• macOS: running python3 --version may offer to install the "command line developer tools". Accept, wait for it to finish, and run the command again. Or install Python from https://www.python.org/downloads/ .
• Linux: most distributions already include python3. If yours does not, install it with your package manager, for example sudo apt install python3 on Ubuntu or Debian, or sudo dnf install python3 on Fedora.

From here on, the course writes python3. On Windows, type py (or python) wherever you see python3.

Step 3. Get the course

On https://github.com/Ani19arya/analytics-from-scratch click the green Code button, then Download ZIP. Unzip it somewhere easy to find, such as your Desktop. (If you already use git, git clone https://github.com/Ani19arya/analytics-from-scratch.git does the same thing.)

Nothing else to install. The course uses only Python and the tools that come with it.

Step 4. Move into the lesson folder

cd means "change directory" (directory is another word for folder). Type cd, a space, then the path:

cd Desktop/analytics-from-scratch/phases/00-start-here/01-run-your-first-lesson

Your folder may be called analytics-from-scratch-main if it came from the ZIP. A shortcut that works on macOS and in most Windows and Linux terminals: type cd with a space, drag the folder from your file window into the terminal, and press Enter.

Three commands help you find your way:

Type ls. You should see README.md, exercise.py, test_lesson.py and the others from the table above.

Step 5. Run the test and watch it fail

python3 test_lesson.py

The exercise is not filled in yet, so the test fails. That is expected. The end of the output looks like this:

======================================================================
FAIL: test_total_on_lesson_data (__main__.TestOrders)
----------------------------------------------------------------------
Traceback (most recent call last):
  File ".../01-run-your-first-lesson/test_lesson.py", line 69, in test_total_on_lesson_data
    self.assertAlmostEqual(got, total, delta=0.01, msg="total_amount gave %r but the amounts add up to %.2f: did the loop add every order?" % (got, total))
AssertionError: 0 != 11194.62 within 0.01 delta (11194.62 difference) : total_amount gave 0 but the amounts add up to 11194.62: did the loop add every order?

----------------------------------------------------------------------
Ran 6 tests in 0.001s

FAILED (failures=4)

How to read it:

1. Each block starting with FAIL: is one check that did not pass. test_total_on_lesson_data is its name, which tells you what it checks.
2. Skip the Traceback and File lines for now. They say where inside the test the check lives.
3. Read the AssertionError line. 0 != 11194.62 means "your code gave 0, the right answer is 11194.62". The text after the colon says what to look at.
4. The last line counts the problems: FAILED (failures=4) means 4 of the 6 checks failed.

Step 6. Fix the code

Open exercise.py in any text editor (Notepad, TextEdit in plain text mode, or a code editor such as VS Code). Look at the first function:

def count_orders(orders):
    """How many orders are in the list?"""
    # TODO: return the number of orders. Hint: len(orders) counts the items in a list.
    return 0

• def count_orders(orders): starts a function: a named recipe that takes something in (here a list of orders) and gives an answer back.
• orders is a list: many values in order, like the rows of a spreadsheet.
• return 0 gives back 0 every time. That is why the test failed.

Replace the TODO and return 0 with return len(orders). len counts the items in a list. Save the file and run the test again. Now it says FAILED (failures=3): one problem fixed.

The second function needs a loop, which does the same step once for each row. This is what the loop does for the first three orders:

from exercise import load_orders
orders = load_orders()
total = 0
for order in orders[:3]:
    total = total + order["amount"]
    print(order["order_id"], order["amount"], round(total, 2))

1001 387.87 387.87
1002 386.78 774.65
1003 397.6 1172.25

orders[:3] means "the first three orders". order["amount"] reads the amount cell of one row. total = total + ... keeps a running total, like adding up a bill line by line. Do the same in total_amount for every order (use orders, not orders[:3]), save, and run the test.

Step 7. Watch it pass

......
----------------------------------------------------------------------
Ran 6 tests in 0.000s

OK

Each dot is a check that passed. OK means all of them did. You can also run your file on its own to see your answers:

python3 exercise.py

orders: 38
total: 11194.62

Step 8. Check the worked answer (optional)

The same test can mark the solution instead of your exercise. On macOS and Linux:

CHECK_SOLUTION=1 python3 test_lesson.py

On Windows PowerShell, set the switch, run, then clear it so later runs check your own file again:

$env:CHECK_SOLUTION=1; py test_lesson.py; Remove-Item Env:CHECK_SOLUTION

Step 9. Learn with the tutor skill (optional)

If you use a coding agent (Claude Code, Cursor, Codex or another agent that reads SKILL.md files), the course has a tutor for it in skills/. Install it as that page explains, open the agent inside your copy of the course, and type "tutor me through analytics from scratch". It asks three questions to pick your starting lesson, then teaches one lesson at a time: it asks you to run the test first, gives a hint before any answer, and only shows solution/ if you ask for it twice. It never edits the test to make it pass. You do not need it: every lesson works with just the README and the test.

Common mistakes

Counting lines instead of orders. A CSV file has a header line with the column names. Count the lines and you get one too many:

with open("data/orders.csv") as f:
    print(len(f.readlines()))

39

The file has 39 lines but 38 orders. Tell Ritu "39" and you have invented an order. load_orders uses csv.DictReader, which reads the header as column names, so len(orders) is 38.

Adding numbers that are still text. A CSV stores everything as text. Joining two pieces of text with + glues them together instead of adding:

print("387.87" + "386.78")

387.87386.78

That is why load_orders turns the amount into a number with float(...) first. Lesson 06 covers this in depth.

Dropping the free orders. Three orders were free replacement meals with an amount of 0.00 (orders 1008, 1011 and 1028). They are still orders. If you skip rows with a zero amount you would report 35 orders, not 38. The total is the same either way, which is why this mistake is easy to miss.

Running the test from the wrong folder. If you run python3 test_lesson.py from the top of the course instead of the lesson folder, Python cannot find the file:

python3: can't open file '.../analytics-from-scratch/test_lesson.py': [Errno 2] No such file or directory

cd into the lesson folder first, and check with pwd.

Typing python on a Mac. On many Macs python does not exist or is an old version. Use python3. On Windows, use py or python.

Your turn

Open exercise.py and fill in count_orders and total_amount. Then run, from this folder:

python3 test_lesson.py

It passes when both functions work on the lesson data, on a second dataset, and on small cases such as an empty list.

Stretch: add a function average_order(orders) that returns the total divided by the count, and decide what it should return for an empty list (you cannot divide by zero). The test does not check it.

Check yourself

1. A test ends with FAILED (failures=3). What do you read first, and what does 0 != 11194.62 mean?
2. Why does each test also run your code on a second dataset?
3. A CSV file has 39 lines. How many rows of data does it have?
4. You see can't open file 'test_lesson.py'. What went wrong?

What to tell the stakeholder

"Yesterday the Pune kitchen made 38 orders worth Rs 11,194.62. Three of the 38 were free replacement meals at Rs 0, so they count as orders but add nothing to the takings."

In real job postings

Of the eight tools the data drop tracks, Python is the most requested after SQL. In the Real Analyst Jobs data drop of 25 October 2026, 996 of the 2,056 US analyst postings that list a tool ask for Python (48.4%), and 448 of 900 in India (49.8%). Browse them at Python data analyst jobs.

Next

02 Think like an analyst