025-weather/docs/business-case.md

Business Case: Weather Data (Day 25)

Metadata

Key	Value
ID	BC-001
CrossReference	SA-001, DICT-001
Language	en
Domain	it

Version History

Date	Status	Author	Reviewer	Change	Commit
2026-10-13	Accepted	Jens Tirsvad Nielsen	AI as stand-in for S01	Initial version, reviewed in RC-001 (Go)	pending

---

Executive Summary

Weather Data (Day 25) is the day-25 assignment of Udemy's 100 Days of Code: The Complete Python Pro Bootcamp: reading comma-separated values (CSV) data in Python with the weather data of one week as the example. The first lecture reads weather_data.csv with Python's built-in csv library, collects the temperatures as whole numbers, and then shows how the pandas library does the same in one line. The second lecture introduces the two data structures of pandas, DataFrame (a table) and Series (a column), converts them to a dictionary and a list, computes the average and the maximum temperature, selects rows by a condition (the row of Monday, the row of the hottest day, Monday's temperature in Fahrenheit), creates a DataFrame from a dictionary, and exports it to a CSV file.

This repository turns both lectures into one small, tested and documented Python package with a command-line entry point, and gives it the identity of a repository: name, README, description, topics, a wiki in English, Danish and Bahasa Melayu, and a screenshot. The work is planned in five milestones, each delivered by one branch and one pull request.

Methodological and Standards Foundation

• Process: the Software Quality Assurance (SQA) and Quality Criteria (QC) framework mounted at framework/: Business Case, Stakeholder Analysis, Project Plan, milestones, tasks synced as issues, then code, each step reviewed before the next.
• Quality model: ISO/IEC 25010:2023; every QC criterion is tagged with a characteristic of it.
• Code: Python Enhancement Proposals (PEP) 8, 257 and 484, reviewed against QC-PY-001; comments in Doxygen style; tests with pytest.
• Domain terms: the lectures' own terms and names, recorded in the Domain Dictionary DICT-001: CSV, row, column, DataFrame, Series, read_csv, to_dict, to_list, mean, max, to_csv.

Problem Statement

• The lectures work in one script with scratch variables; a participant who wants to compare solutions or re-run a step has to edit the script.
• Reading a CSV file with the csv library is easy to get subtly wrong: the header row is data to the reader, every value is a string, and the file path depends on the current working directory.
• The pandas library is a large dependency; a repository that needs it to run any code at all is heavy for a visitor who only wants to see the csv part.
• A repository without a description, topics, README, wiki or a picture of the result is hard to find and to judge for the people who browse for ideas.

Business Opportunity

A finished, reproducible repository shows the path from the day-25 lectures to documented, tested code and lets the csv way and the pandas way be compared side by side on the same file. It can be shared with, and compared against, the solutions of other participants.

Objectives

#	Objective
1	Provide the weather data of the lecture: weather_data.csv with the columns day, temp and condition for Monday to Sunday, kept in the repository
2	Deliver the csv part of the lecture: read the file with the built-in csv library into a list of rows and extract the temperatures as whole numbers, leaving out the header row
3	Deliver the pandas reading of the lecture: read the file with pandas.read_csv, and show that the result is a DataFrame and that a column of it is a Series
4	Deliver the conversions of the lecture: DataFrame.to_dict() and Series.to_list()
5	Deliver the computations of the lecture: the average temperature by sum and length and by Series.mean(), and the maximum temperature by Series.max()
6	Deliver the row selection of the lecture: the row of a given day (Monday), the row of the hottest day, and the temperature of a day converted to Fahrenheit (temp * 9 / 5 + 32)
7	Deliver the creation and export of the lecture: build a DataFrame from a dictionary (students and scores) and write it to a CSV file with DataFrame.to_csv
8	Keep the code simple to read and the names of the lectures (data, temperatures, temp_list, data_dict, average_temp, monday, to_dict, to_list, mean, max, to_csv), constants in constants.py
9	Provide a reproducible environment: Python 3.13 or newer, a local .venv, a pyproject.toml; the csv part runs with zero runtime dependencies and pandas is an optional extra
10	Prove the behaviour with pytest tests that use temporary files and never overwrite a file of the user
11	Document the project: Doxygen comments in the source with a Doxyfile, and a README with set-up instructions for Windows PowerShell, Linux Debian and macOS
12	Keep every step traceable: one branch and one pull request per milestone, each pull request closing the issues it completes
13	Publish the repository with a description and topics, and a wiki in English, Danish and Bahasa Melayu
14	Show the result in a real screenshot of the program running in a terminal, kept in the repository and shown in the README and on the wiki

Scope

In Scope

• The two day-25 lectures on reading CSV data and on DataFrame and Series, as listed in the objectives.
• The data file weather_data.csv that S01 downloaded for the lecture, adopted unchanged.
• Constants in constants.py; source in src/, tests in tests/, documents in docs/.
• pyproject.toml, Python .gitignore, Doxyfile, README.md following the Product Owner's template, a continuous integration (CI) workflow.
• Repository description and topics on the git host.
• A wiki in English, Danish and Bahasa Melayu, in the form of the wiki of TirSystem/github-action.
• A screenshot of the program running in a terminal, kept in docs/images/ and shown in the README and on the three wiki pages.

Out of Scope

• The later pandas lessons the lecturer announces (NumPy, Matplotlib, plots, aggregation beyond the lectures).
• Reading from the internet, a database or a spreadsheet; creating the Google Sheets document of the lecture.
• A graphical user interface.
• Packaging for or publishing to the Python Package Index (PyPI).
• A tool in the repository that takes the screenshot: the screenshot is taken once with a one-off script outside the repository.
• Using or testing the tokens in .env; the file is for personal use only and is never imported by the project.

Expected Benefits

Tangible Benefits

• A finished day-25 solution that runs with one command after the README steps, and whose csv part needs nothing to install.
• A test suite that runs in continuous integration and never touches a file of the user.
• Generated source documentation.
• A repository page with description, topics, README, a three-language wiki and a screenshot.

Intangible Benefits

• Practice with the csv library and the basic pandas structures, which is the aim of the lectures.
• A pattern for the following days of the course.
• Confidence from a step-by-step, reviewed delivery.

Strategic Alignment

The project supports the participant's goal of finishing the bootcamp with repositories that other people can read, run and learn from, and it exercises the framework's rule that planning, review and code stay in step.

Success Criteria

#	Criterion	Target	Measure
1	The data file is the lecture's	weather_data.csv has the header day,temp,condition and 7 data rows, Monday to Sunday	Test of the file; review
2	The csv reader is correct	Temperatures read as [12, 14, 15, 14, 21, 22, 24], all int, header skipped	Matching tests
3	The pandas reader is correct	read_csv returns a DataFrame with 7 rows; its temp column is a Series; to_list gives the same temperatures as the csv reader	Matching tests
4	The computations are correct	Average is 17.428... (the manual and the mean way agree), maximum is 24	Matching tests
5	The row selection is correct	The Monday row has Sunny and 12; the hottest day is Sunday; Monday in Fahrenheit is 53.6	Matching tests
6	Create and export work	A DataFrame made from the students/scores dictionary is written to a CSV file and reads back equal	Matching tests with tmp_path
7	Tests are green	100% of tests pass; 0 tests write outside a temporary directory	pytest exit code 0 locally and in CI
8	The set-up steps work	A fresh clone reaches a green pytest by following the README alone, on Windows PowerShell (by S01) and on Linux (by CI)	One run per operating system, recorded in the pull request
9	Source documentation builds	doxygen Doxyfile ends with 0 warnings	Doxygen output
10	The csv part needs no dependency	[project].dependencies is empty and the csv reader imports without pandas installed	pyproject.toml; a test of the import
11	Steps are traceable	5 of 5 milestones merged by pull request, each description with one Closes #N line per completed issue	Pull request list and issue states
12	The repository page is complete	Non-empty description, at least 5 topics, and the wiki pages Home, Home.da and Home.ms	Git host
13	The result is shown in a real screenshot	docs/images/weather-run.png is a capture of a terminal that runs the program; the README and the three wiki pages show it, and its address on main answers HTTP 200 with an image	The file, the README and the wiki, checked against the Go/No-Go list of MIL-005

Risks

Risk	Impact	Mitigation
pandas is a heavy dependency and has to support Python 3.13	Installation fails or is slow	pandas is an optional extra, the csv part runs without it; the pin is a lower bound only and is checked in CI
Debian 12 ships Python 3.11, below the required 3.13	Linux set-up fails	README states Debian 13 or newer, or a separately installed Python 3.13
The data file path depends on the working directory	The program finds no file when started elsewhere	The default path is resolved against the package folder, not the working directory; the path is a parameter
A test overwrites a file of the user	The user loses data	The export takes a path; tests pass a temporary path (pytest tmp_path)
Author and reviewer are the same person (S01)	A defect can pass review unnoticed	Review against the QC checklists, record each review as an RC-*, and let the pull request be the second look
The lectures are available only as the summaries in the request	The code differs from the video	S01 compares the finished program with the video at the MIL-003 and MIL-004 Go/No-Go
Tokens in .env leak into the repository	Credentials exposed	.env is in .gitignore, is never imported, and is not part of any task
The wiki publish workflow needs a secret of the organization that may be missing	The wiki stays empty	The pages are files in wiki/ of the pull request; a missing secret is reported, not worked around
The screenshot is a capture of a real terminal window, which needs a display	It cannot be repeated in CI	It is taken once on S01's machine with a one-off script outside the repository and committed as a file

Assumptions

• Python 3.13 or newer is available on the machines that run the program.
• S01 is the only person who reviews and accepts artifacts.
• The git host is the Tirsvad Gitea instance, and its issues and milestones are used for tracking.
• The two lecture summaries in the request are the specification of day 25; the video is the tie-breaker when a detail is missing.
• weather_data.csv from S01's download folder holds the data of the lecture and is adopted unchanged.
• The wiki of the repository is enabled and has no page yet.

Constraints

• Python 3.13 or newer, in a virtual environment (venv).
• pytest for tests; constants.py for constants; pyproject.toml for project configuration; a Python .gitignore; Doxygen comments in the source and a Doxyfile.
• Folder structure src/, tests/, docs/.
• No runtime dependencies unless needed; pandas is needed only for the second half of the lectures and is therefore an extra.
• Plan window from 2026-10-14 to 2026-10-23 (proposed, see PP-001).
• S01 switched on autonomous mode in chat on 2026-10-13: S01's acceptance of the plan is the request to commit, push, open and merge the pull requests; the assistant's reviews are then not independent (see AGENTS.md, Modes).
• The plan gate holds: no file under src/ or tests/ before an accepted, reviewed milestone that lists the task.

Cost–Benefit Assessment

The assessment is qualitative on purpose: this is an unpaid learning project with one participant, so money does not measure either side.

Costs	Benefits
About eight days of S01's spare time, most of it review of the planning documents	A finished and shareable repository (objectives 1 to 14)
pandas, Doxygen and pytest as tools	A comparison of the csv and the pandas way, source documentation and automatic checks

Stakeholders

Stakeholder ID (SA)	Interest in this project
S01	Wants a correct, tested and documented solution of the assignment: objectives 1 to 14
S02	Needs readable, runnable code with the lectures' names and README instructions (objectives 2 to 9 and 11)
S03	Needs a clear repository page and as little as possible to install (objectives 9, 11, 13 and 14)

Recommendation

Proceed — the work is small (two lectures, a handful of functions, tests, documentation and publishing), the cost is about eight days of the Product Owner's spare time, and the result is a reusable, documented repository.

---