From Python Foundation to ML workflows

First complete the relevant Python Foundation lessons. This bridge adds skills the introductory course intentionally does not require. Allow additional guided practice; completing the introductory course alone does not establish dataframe or file-I/O mastery.

Records and collections

record = {"category": "network", "minutes": 15}
records = [record, {"category": "software", "minutes": 20}]
print(record["minutes"])
print(sum(item["minutes"] for item in records))

Predict 15 and 35 before running. A dictionary describes one record; the list contains two records. The generator supplies each duration to sum. Rewrite it as the explicit loop you learned in Foundation. Then add a third record and predict the new total.

Imports, folders and files

Use workspace setup first. Run from the folder containing this repository's README, pyproject.toml and data folder.

from pathlib import Path
import csv
import json
root = Path.cwd()
print("Working folder:", root)
path = root / "data/samples/service-files/tickets.csv"
with path.open(newline="", encoding="utf-8") as stream:
    tickets = list(csv.DictReader(stream))
minutes = [int(row["minutes"]) for row in tickets]
print(len(tickets), sum(minutes))
output = root / "runs/bridge"
output.mkdir(parents=True, exist_ok=True)
(output / "summary.json").write_text(json.dumps({"minutes": sum(minutes)}), encoding="utf-8")

Expect four rows and 60 minutes. CSV fields arrive as text; int converts them before arithmetic. A with block closes the file even if processing fails. pathlib and csv are standard-library modules; pandas is installed separately. FileNotFoundError suggests inspecting the working folder first, not reinstalling Python.

Dataframes and packages

After the core installation in first run:

import pandas as pd
frame = pd.read_csv(path)
print(frame[["category", "minutes"]].head())
print(frame.groupby("category")["minutes"].sum())

A dataframe is a table; selecting a column returns a series. Grouping collects rows by a key before aggregation. Compare the grand total with your standard-library result. A table library reduces repeated code, but does not choose the right unit or missing-value policy for you.

Independent completion check

On a copy, add a valid ticket and reconcile the total in both implementations. Replace one duration with unknown; explain the conversion error and choose a visible validation policy. Do not silently substitute zero. Save your modified code, input copy, expected total, observed total and one error explanation. You are ready for Module 2 when you can repeat this with changed records and explain the imports and paths without the completed example.