Capstone: From Raw Rows to Insight Report

18 min 45 XP

Study loop Read Predict Run Tweak Prove Review

Lesson 5 of 5 · View course roadmap

Step 1

Learn the idea

Time to run the full pipeline every analyst runs, whatever the tool:

Parse — split each "region,amount" row into fields
Clean — convert amounts, quarantine garbage with try/except
Aggregate — total by region with the dict accumulator
Report — row count, grand total, top region

Parse → clean → aggregate → report is the same skeleton whether the data is 6 rows in a list or 6 billion in a warehouse; only the tools scale up. Ship this and you've done real analysis — the kind that answers an actual business question.

Where you'll use this

This is the nightly job at thousands of companies: parse yesterday's transactions, skip the garbage, total by region, email the summary. You've written the core of a BI pipeline in ~15 lines.

Common mistakes

Counting a row as valid before it survives conversion — increment inside the try, after int() succeeds.
row.split(",") assuming exactly two fields — a stray comma shifts everything; in real work check the field count.
Reporting totals but not the denominator — '490 revenue' means little without 'from 5 valid rows of 6'.

Pro tip

Structure real pipelines as three functions — parse(rows), aggregate(records), report(totals) — so each stage is testable alone. The csv module then replaces your split(",") for free.

Step 2

Try it yourself

Blank · autosaved

The lesson example is loaded and ready — press Run, then change something and run it again. Breaking it is part of learning. Want a clean slate? Tap “New blank”.

PYexample.py

⌘ + Enter to run

Output appears here…

Step 3

Pass the challenge +45 XP

Blank · autosaved

Analyse rows = ["north,120", "south,95", "north,80", "west,x", "east,150", "south,45"], skipping rows whose amount isn't a number. Print three lines: the number of valid rows, the grand total, and the region with the highest total.

Target output

5
490
north

PYchallenge.py

rows = ["north,120", "south,95", "north,80", "west,x", "east,150", "south,45"]
totals = {}
valid = 0
for row in rows:
    region, raw_amount = row.split(",")
    try:
        amount = int(raw_amount)
    except ValueError:
        continue
    valid += 1
    totals[region] = totals.get(region, 0) + amount
print(valid)
print(sum(totals.values()))
print(max(totals, key=totals.get))

Run your code to check it…

Step 4

Check your understanding

1. Put the analysis pipeline in order: report → parse → clean → aggregate parse → clean → aggregate → report aggregate → parse → report → clean clean → report → parse → aggregate

2. Why 'continue' on a bad row instead of letting it crash? continue is faster Real datasets always contain garbage; the report must still ship Python requires it To keep the row for later

Last step

Your notes (saved on this device)

Tip: use ← and → to move between lessons, ⌘K to search everything.