> For the complete documentation index, see [llms.txt](https://opencampus.gitbook.io/opencampus-machine-learning-program/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://opencampus.gitbook.io/opencampus-machine-learning-program/courses/einfuehrung/week-04-git-2-and-feature-engineering.md).

# Week 4 - Versioning with Git (Part 2) and Feature Engineering

**Question of the week (from the operations manager):** *What do we know the day before that could predict tomorrow’s sales?*

**Canvas focus:** blocks 6 and 7 (Known in advance, Data decisions) · **Skills:** `leakage`, `/to-analysis-plan`

### This week you will ...

* collaborate in a team repository: push, pull, resolve conflicts, and commit only what passes the checks,
* encode categorical variables with one-hot encoding, and explain embeddings as learned dense representations,
* distinguish features known in advance from features only known afterwards (a first look at data leakage),
* apply feature engineering: binning, joining categories, interaction effects.

### Until next session you should ...

* [x] **Create the three features** your team chose and add them to the data preparation folder of your repository.<br>
* [x] **Analyse the time series** of the six product groups: moving averages, percent change, seasonality.<br>
* [x] **Create a** [**Kaggle**](https://www.kaggle.com/) **account** and review the [course competition](https://www.kaggle.com/t/d832497c95d744de915ad9eb80ad9ed6). Find out what it fixes for your project and record it where your agent will see it. It scores forecasts for a whole year without that year’s sales, so features based on recent sales cannot be used for the submission.<br>
* [x] **Check on EduHub** that your team is confirmed and every member is listed as an author.<br>
* [x] **Agree on checks before code:** run `/to-analysis-plan` and confirm the checkpoints. For each check, name a source of truth that the pipeline does not compute itself, and replace any check that would pass by construction.<br>
* [x] **Watch the videos and do the readings** listed below (≈ 81 minutes).

**Self-study:** lesson 4 of [Learn it in seven lessons](https://github.com/steffen74/agentic-data-science/blob/main/docs/lessons.md) covers `/to-analysis-plan`.

| Lesson / material                                                                                                                                                          | Source                              | Min    |
| -------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------- | ------ |
| [DateTimes](https://learn.deeplearning.ai/specializations/data-analytics/lesson/yt06e/datetimes)                                                                           | Python for Data Analytics, M5       | 5      |
| [Using DateTimes as Indices](https://learn.deeplearning.ai/specializations/data-analytics/lesson/xtibp/using-datetimes-as-indices)                                         | Python for Data Analytics, M5       | 3      |
| [Moving Averages](https://learn.deeplearning.ai/specializations/data-analytics/lesson/n5vs5/moving-average)                                                                | Python for Data Analytics, M5       | 5      |
| [Percent Change](https://learn.deeplearning.ai/specializations/data-analytics/lesson/icu5f/percent-change)                                                                 | Python for Data Analytics, M5       | 4      |
| [Segmentation](https://learn.deeplearning.ai/specializations/data-analytics/lesson/n5s9q/segmentation)                                                                     | Python for Data Analytics, M5       | 5      |
| [Encoding categorical data](https://learn.deeplearning.ai/specializations/data-analytics/lesson/htxtl/encoding-categorical-data)                                           | Python for Data Analytics, M4       | 3      |
| [Modeling with categorical data](https://learn.deeplearning.ai/specializations/data-analytics/lesson/r7nt2/modeling-with-categorical-data)                                 | Python for Data Analytics, M4       | 5      |
| [Categorical data: one-hot encoding](https://developers.google.com/machine-learning/crash-course/categorical-data/one-hot-encoding) (reading)                              | Google ML Crash Course              | 15     |
| [Embeddings](https://developers.google.com/machine-learning/crash-course/embeddings), first pages (reading)                                                                | Google ML Crash Course              | 15     |
| [Feature specification](https://learn.deeplearning.ai/courses/spec-driven-development-with-coding-agents/lesson/15v9p3/feature-specification) (before `/to-analysis-plan`) | Spec-Driven Development course      | 3      |
| [Feature implementation](https://learn.deeplearning.ai/courses/spec-driven-development-with-coding-agents/lesson/eg5lav/feature-implementation)                            | Spec-Driven Development course      | 1      |
| [Feature validation](https://learn.deeplearning.ai/courses/spec-driven-development-with-coding-agents/lesson/0kj4sj/feature-validation)                                    | Spec-Driven Development course      | 4      |
| [Word Representation](https://learn.deeplearning.ai/specializations/deep-learning/lesson/pqt4m/word-representation)                                                        | Deep Learning Specialization, C5 W2 | 10     |
| [Embedding Matrix](https://learn.deeplearning.ai/specializations/deep-learning/lesson/r4twj/embedding-matrix)                                                              | Deep Learning Specialization, C5 W2 | 3      |
| **Total**                                                                                                                                                                  |                                     | **81** |

{% hint style="info" %}
**Deliverable for the operations manager:** three findings about trends and seasonality, each with one chart and one sentence on what it means for production planning.
{% endhint %}

#### Optional

* Spec-Driven Development course: Project replanning (6 min), The second feature phase (6 min), The MVP (3 min).
* Deep Learning Specialization, C5 W2: Using Word Embeddings (9 min), Properties of Word Embeddings (11 min).
* The first two lessons of the short course “Embedding Models: from Architecture to Implementation” (DeepLearning.AI).

**Estimated workload:** about 11.6 hours (session 1.75 h, videos and readings 1.8 h, tasks 8 h)


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://opencampus.gitbook.io/opencampus-machine-learning-program/courses/einfuehrung/week-04-git-2-and-feature-engineering.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.