> For the complete documentation index, see [llms.txt](https://www.mlcompendium.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://www.mlcompendium.com/data/processing-intro.md). # Data Processing Raw values are rarely on the scale or in the shape a learner expects. After this chapter the reader can scale, transform, impute, and annotate features, and can point at a pipeline that does it. The chapter starts with shape, because skewed data breaks tests that assume a normal shape: [Distribution Transformation](/data/processing-intro/distribution-transformation.md) transforms the column, mainly with Box-Cox, and then checks whether normality actually arrived. Shape is only half of it, since features on different scales pull a model toward the loud columns, and [Normalization & Scaling](/data/processing-intro/normalization-and-scaling.md) puts them on a shared scale without inventing signal. Those steps have to run somewhere repeatable, so [Data Pipelines](/data/processing-intro/data-pipelines.md) moves data from source systems into a place a model or warehouse can use and asks which integration path and orchestrator own that path. With clean columns flowing, [Feature Engineering](/data/processing-intro/feature-engineering.md) builds features, fills gaps, and measures how alike two rows are. The last input is the label itself, and [Annotation & Disagreement](/data/processing-intro/annotation-and-disagreement.md) treats labels as work product that annotators disagree on, from myths and agreement metrics to the tools that run that loop. To see this step on a real table, [Cleaning Data With AI Denoisers](https://pub.towardsai.net/cleaning-data-with-ai-denoisers-be1bdea0fe20) (October 2024) shows this step on a real table: it presents AI denoisers as algorithms that reduce noise in data such as images, audio, and video, where traditional denoising relies on filters and statistical methods. --- # 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://www.mlcompendium.com/data/processing-intro.md?ask=&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.