ETL (Extract, Transform, Load) The process of pulling data from various sources, cleaning it, and loading it into a destination. Example: Apache Airflow, dbt (data build tool).

Parquet A columnar storage file format that is highly optimized for reading large datasets efficiently. Context: Used in "Data Lakes" instead of CSVs for speed.

Data Pipeline A series of automated steps that move data from a source to a destination. Example: Pulling daily sales from Shopify ⇒ Cleaning ⇒ Updating a Dashboard.

SQL (Structured Query Language) The standard language used to communicate with relational databases. Example: PostgreSQL, MySQL, SQLite.

CI/CD (Continuous Integration / Deployment) A set of practices that automate the testing and deployment of code, ensuring that updates to data pipelines are reliable and seamless. Example: GitHub Actions, GitLab CI.

Hypervisor Software that creates and runs virtual machines (VMs), allowing one physical server to host multiple isolated operating systems. Example: KVM (Kernel-based Virtual Machine), LXD/Incus (container hypervisor).

Predictive Analytics Using historical data to make predictions about future events. Example: Predicting customer churn based on usage patterns.

Pandas The gold-standard Python library for data manipulation and analysis. Other options here include Polars, along with Narwhals as abstraction layer. Product: Open-source Python library.

Clustering An unsupervised learning technique that groups similar data points together. Context: Customer segmentation for targeted marketing.

Scikit-Learn A robust open-source library for classical machine learning algorithms. Product: Used for regression, classification, and clustering.

RandomForest An ensemble learning method that constructs multiple decision trees to improve prediction accuracy and prevent overfitting. Example: Used in Scikit-Learn for complex classification tasks.

Box Plot A standardized way of displaying the distribution of data based on a five-number summary, excellent for visualizing variance and outliers. Context: Identifying "extreme" values in delivery times.

Correlation Matrix A table showing correlation coefficients between variables, helping analysts understand how different business metrics move together. Example: Visualized via Seaborn or Plotly heatmaps in Python.

LLM (Large Language Model) AI trained on vast amounts of text to understand and generate human-like language. Example: GPT-4, Llama 3, Gemma 4, Mistral.

Transformer The neural network architecture that makes modern LLMs possible by "paying attention" to different parts of a sequence. Library: Hugging Face Transformers.

PyTorch / TensorFlow The two primary frameworks used to build and train deep learning models. Product: Open-source libraries by Meta (PyTorch) and Google (TensorFlow).

API (Application Programming Interface) A bridge that allows two pieces of software to talk to each other. Context: Sending data from your app to an AI model via a REST API to get a forecast.

Neural Network Computing systems inspired by biological brains that learn patterns through layers of interconnected nodes. Example: The foundation of Image Recognition and Deep Learning.

MCP (Model Context Protocol) An open standard that enables AI models to seamlessly connect to diverse data sources and tools without custom glue-code for every integration. Example: Anthropic's MCP for connecting LLMs to local databases.