# Glossary

* [A](/products/glossary#h3_372029405)
* [B](/products/glossary#h3_372029408)
* [C](/products/glossary#h3_372029407)
* [D](/products/glossary#h3_372029402)
* [E](/products/glossary#h3_372029401)
* [F](/products/glossary#h3_372029404)
* [G](/products/glossary#h3_372029403)
* [H](/products/glossary#h3_372029414)
* [I](/products/glossary#h3_372029413)
* [J](/products/glossary#h3_372029416)
* [M](/products/glossary#h3_372029409)
* [N](/products/glossary#h3_372029412)
* [P](/products/glossary#h3_372029422)
* [R](/products/glossary#h3_372029424)
* [S](/products/glossary#h3_372029423)
* [T](/products/glossary#h3_372029418)
* [U](/products/glossary#h3_372029417)
* [W](/products/glossary#h3_372029419)


### A

**activation**

An activation allows marketers to personalize communication to users who might be interested in their brand.

**AI agent**

An AI agent is an intelligent assistant that uses artificial intelligence to help users create and refine segments, analyze audiences, and generate insights. AI agents can understand natural language queries, provide recommendations, and automate complex tasks that would typically require manual analysis.

**AI Agent Foundry**

A development environment within Treasure Data that enables users to build, train, and deploy custom AI agents. It provides tools and resources for creating specialized agents tailored to specific business needs and use cases.

**ANSI SQL**

A standard for SQL specified by the American National Standards Institute (ANSI).

**authentication**

A security credential configuration that stores connection information and access tokens for external services and data sources. Authentications enable Treasure Data to securely connect to third-party platforms (such as databases, cloud storage, SaaS applications, and marketing tools) for data import and export operations. Multiple integrations can reuse the same authentication, and authentications can be managed centrally through the Integration Hub.

**Atomicity**

In database systems, atomicity is one of the ACID (*Atomicity,*[*Consistency*](https://en.wikipedia.org/wiki/Consistency_(database_systems)) *,*[*Isolation*](https://en.wikipedia.org/wiki/Isolation_(database_systems)) *,*[*Durability*](https://en.wikipedia.org/wiki/Durability_(database_systems))) transaction properties. An atomic transaction is a series of database operations that cannot be divided or reduced such that one of the following happens:

* all of the transaction occurs
* none of the transaction occurs


This ensures that databases or tables will not be partially updated in certain cases which might create problems. This also means that the progress of the transaction cannot be observed by another database client. It shows either as  *not occurred* or as  *completed*.

In Treasure Data, the four supported modes below are **atomic** because they use a temporary table to store the incoming data before attempting to modify the destination table:

* Append
* Replace
* Truncate
* Update


**Audience Studio**

Treasure Data's intuitive Console interface which allows you to gain deeper insights into your customers more quickly and efficiently.

**AutoML**

AutoML stands for Automated Machine Learning, an established industry practice that automates certain aspects of the end-to-end Machine Learning process. This broadens the user base beyond ML experts to take advantage of machine learning, enabling the development of high-quality Machine Learning models to address a wide range of business needs.

**AutoML Base Package**

The AutoML base package provides generic AutoGluon notebook that may need to be configured to meet specific use cases. It provides general ML training, prediction, and dataset loading operators.

**AutoML Model Sharing**

It is possible to enable sharing of an AutoML model. This enables one user to train the model, and then for a different user to utilize the generated model on their data to make predictions.

**Avro**

A schema-based serialization system that is a preferred format for Hadoop. The serialized data is binary, and is accompanied by the schema that allows the binary data to be de-serialized.

### B

**batch segment**

A collection of audiences that are based on attribute and behavior criteria for activation and reporting.

### C

**catalog**

Treasure Data's collection of schema-flexible integrations designed to work your existing technology stack so you can unlock the power of customer data to deliver stellar brand experiences at scale.

**CCPA**

California Consumer Privacy Act: Regulations governing the how the personal information of California consumers is collected and processed.

**Cookie Consent**

A Treasure Data product that creates and manages website banners that customers can display on their websites to collect consent for cookies from their end users.

**cron**

A time-based job scheduler that runs tasks at fixed times, dates, or intervals.

### D

**DAG**

A directed acyclic graph (DAG) is a conceptual representation of a series of activities. The order of the activities is depicted by a graph, which is visually presented as a set of circles, each one representing an activity, some of which are connected by lines, which represent the flow from one activity to another. Each circle is known as a “vertex” and each line is known as an “edge.” “Directed” means that each edge has a defined direction, so each edge necessarily represents a single directional flow from one vertex to another. “Acyclic” means that there are no loops (i.e., “cycles”) in the graph, so that for any given vertex, if you follow an edge that connects that vertex to another, there is no path in the graph to get back to that initial vertex.

**data enrichment**

The process of merging and adding first-party and third-party data to a dataset.

**data model**

Data models define the structure of your data. It defines how elements and attributes in your database are used to describe customers.

**Data Tanks**

Data Tanks are PostgreSQL databases that are used to accelerate analytical queries.

**database**

An accessible collection of data designed to be analyzed, managed, and updated.

**decision point**

An event that splits the journey into multiple journey branches based on a profile's activity or attributes.

**Digdag**

A tool that allow you to build, run, schedule, and monitor complex workflows.

**entry criteria**

A condition a profile must meet to enter a journey stage.

### E

**ETL**

ETL is a process that extracts, transforms, and loads data from multiple sources to a data warehouse or other unified data repository.

**event**

Customer data and behavior that can be tracked and assessed for reliable analytics and targeted activations.

### F

**folder**

Folders are containers in which you can organize information.

**Foundry Workspace**

A collaborative development environment within AI Agent Foundry where users can create, test, and iterate on AI agents. The workspace provides tools for agent configuration, testing, and deployment management.

**Foreign Key**

Foreign keys enforce referential integrity by ensuring data conforms to certain rules when it is added to a database.

**funnel**

A funnel is simple workflow with stages in which profiles progress. You can configure each stage.

### G

**GDPR**

General Data Protection Regulation. Rules governing privacy and the processing of personal data for persons residing in countries that are part of the European Union (EU).

**gem**

RubyGems is a package manager for the Ruby programming language. A Ruby Gem is an installable package containing Ruby code.

**geolocation database**

A database that contains geographical and other information about IP addresses. Treasure Data's geolocation features are based on MaxMind.

**Gluon Predict**

This AutoML base package notebook makes predictions using a prediction model built by gluon train AutoML notebook.

**Gluon Train**

This is a base package AutoML notebook that builds a prediction model using AutoGluon. It can automate a data scientist's daily operations such as data processing, feature engineering, model selection, ensembling, and hyperparameter tuning.

**GTM**

Google Tag Manager

### H

**Hive**

An SQL query engine that is geared toward processing large databases. Hive queries can complete even with failures.

**Hivemall**

Hivemall is a scalable machine learning library for use with Hive.

### I

**interface**

A user interface (UI) or application programming interface (API) that enables interaction with a system or service. In Treasure Data, the Console provides visual interfaces for various products. The AI Agent Foundry interface, for example, provides a workspace environment where users can build, configure, test, and deploy custom AI agents through an intuitive visual interface.

**integrations**

Schema-flexible connectors designed to work with your existing technology stack to ingest all types of data.

### J

**JDBC driver**

A software component that enables applications to connect to and query Treasure Data databases using standard SQL.

**journey**

A customer journey is a multi-stage orchestration workflow that identifies and progresses customer profiles through different stages of the buying cycle (awareness, consideration, purchase, retention) based on their behaviors and attributes. Journeys use a drag-and-drop canvas to connect campaigns across channels and business units, enabling marketers to create omnichannel experiences that adapt to individual customer behaviors. Journeys can be batch (processing profiles incrementally on a schedule) or real-time (triggering immediately based on events).

**Join Key**

Join keys represent shared dimensions between data sources.

### M

**Matched activations**

Matched activations are all sequential-enabled activations that share the same authentication.

**maven**

Maven is a java tool that manages file versions and file dependencies in the build process.

**merge**

Consolidate all decision paths into a single path after a series of messages/activations are completed.

**Multi-Touch Attribution**

Multi-touch attribution (MTA) is a method of marketing measurement, determining the value of each customer touchpoint that leads to a conversion. MTA helps marketers to figure out which marketing channels or campaigns should be credited with the conversion. Marketers can optimize marketing channels using the MTA AutoML solution notebook.

**Master Key**

An API key that provides full access to perform all permitted operations in Treasure Data based on the user's permission level and access, including both read and write operations. Master keys should be kept secure and not shared publicly, as they grant complete access to your Treasure Data account and data.

### N

**Nested Workflow**

A user-defined workflow that is configured to be triggered by an activation. It processes the output of the activation before the output data is exported, that is, activated. It is an integral part of the activation and the activation would have successful status only when both the activation and nested workflow are successful.

**Next Best Action**

Next-best Action (NBA) is a strategy that helps businesses identify the most effective marketing actions to take to drive customers closer to a desired conversion event. It is designed to optimize marketing efforts and improve the return on investment (ROI) of marketing campaigns. NBA can be implemented using an AutoML solution notebook.

### P

**PII**

Personal Identifiable Information

**pip**

A python-based package manager used to install software.

**PostgreSQL**

PostgreSQL, also known as Postgres, is a free, open-source relational database management system.

**predictive scoring**

Predictive scoring is data-driven lead scoring methodology that uses historical and activity data and predictive modeling to predict behavior.

**Presto**

A distributed SQL query engine that is useful for processing medium amounts of data. Presto queries do not tolerate failures.

**profile**

A profile reflects data about a customer that might include any combination of behaviors, characteristics, and attributes that help to further determine how you want to segment the customers or accounts.

**Parent Segment**

A unified data model in Data Workbench (also called Master Segment) that combines master tables with attribute and behavior tables to create a comprehensive view of customer profiles. Parent segments serve as the foundation for creating more targeted segments in Audience Studio and are refreshed on a scheduled basis.

### R

**result export**

A data export operation that sends query results or processed data from Treasure Data to external destinations such as databases, cloud storage services, data warehouses, or business intelligence tools. Result exports can be configured as one-time operations or scheduled to run automatically, and they support various output formats and authentication methods to integrate with different target systems.

**REST API**

A REST API (also known as RESTful API) is an application programming interface that conforms to the constraints of REST architectural style. A RESTful endpoint is a URL.

**Ruby Gem**

RubyGems is a package manager for the Ruby programming language. A Ruby Gem is an installable package containing Ruby code.

**Rubygems**

RubyGems is a package manager for the Ruby programming language. A Ruby Gem is an installable package containing Ruby code.

### S

**segment**

A segment or segments describe a collection of customer account profiles. Profiles might reflect any category of customers with similar likes, dislikes, or demographic data.

**source**

A data source configuration that defines where and how to import data into Treasure Data from external systems. Sources specify connection details, data formats, import schedules, and transformation rules for ingesting data from various platforms such as databases, cloud storage, APIs, SaaS applications, streaming services, and file uploads. Sources are managed through the Integration Hub and can be configured for both batch and streaming data ingestion.

**segments**

Segments describe a collection of customer account profiles.

**Sequential Activation**

You can enable Sequential Activation on an activation with Nested Workflow. Any sequential-enabled activation does not run in parallel with matched activations. Matched activations are all sequential-enabled activations that share the same authentication.

**SHAP Analysis**

SHAP stands for SHapley Additive exPlanations. SHAP Analysis refers to a custom solution AutoML notebook. Treasure Data has developed a multi-touch attribution (MTA) model that combines Long Short-Term Memory (LSTM) deep learning with Shapley values, a concept developed by Nobel Prize-winning economist Lloyd Shapley. The Shapley Values model extracts meaningful statistical patterns from user journey data imported into Treasure Data’s Customer Data Platform (CDP).

**SQL**

Structured Query Language

**stage**

A stage describes how a user advances along their path toward a purchase. Treasure Data Journey Orchestration supports up to eight stages. After marketers create journey stages, they can further analyze and refine the stages and activate specific stages for campaigns.

**Stale Profile**

The state of a profile which has not completed the journey goal for a stage after "x" amount of time.

Stale profiles are captured as a segment or attribute that marketers can reference when redirecting profiles, building ordinary segments, or using for reporting.

### T

**tags**

Treasure Data's tags allow you to provide additional security surrounding access to private or sensitive data. Tags describe columns that contain a specific type of information (for example, personal identifiable information, finance, sensitive, and so on).

**TD Toolbelt**

TD Toolbelt allows you to use your command-line interface to access features of Treasure Data.

**Time Series Forecasting**

Time series forecasting is the task of fitting a model to historical, time-stamped data to predict future values. This AutoML notebook trains time-series models and forecasts future values.

### U

**UDF**

User-defined functions.

**UDFs**

User-defined functions.

**Unix Time**

The number of seconds that have elapsed since January 1, 1970. An example of a Unix Time stamp is 1672338261.

**utilization dashboard**

A visualization showing how much Treasure Data resources are used. Allowing you to monitor progress and assess productivity.

**UUID**

Universal Unique Identifier (UUID) is used to uniquely label an object or entity on the Internet. It is a 128 bit value that is unique for all practical purposes when generated by standard methods.

### W

**wait step**

If the profile hasn't completed the journey goal, wait "X" time before progressing to the next step of the journey.

**Workflow**

Treasure Workflow is a multi-cloud task engine that can orchestrate tasks not only on Treasure Data but across a variety of cloud infrastructures such as AWS and Google Cloud Platform. TD Workflow is based on Digdag. Most, but not all, Digdag operators can be used as part of Treasure Workflow.

**Write-only Key**

An API key that provides a restricted level of access, allowing only data import operations to Treasure Data databases that the user has write access to. Write-only keys are useful for providing access to third parties or embedding in ingestion libraries (such as web page integrations) without exposing full account access. Unlike Master keys, Write-only keys cannot read or delete data.