ABAC (Attribute-Based Access Control)
ABAC decides access by evaluating attributes of the person, the resource, the action, and the context against a policy, instead of by member...
Read definitionExplore our comprehensive data dictionary to understand key data concepts, definitions, and relationships
ABAC decides access by evaluating attributes of the person, the resource, the action, and the context against a policy, instead of by member...
Read definitionACID transactions are the four guarantees that keep database changes correct: atomicity, consistency, isolation, and durability. They make s...
Read definitionThe AI Act is the European Union regulation that governs artificial intelligence. It sorts AI systems by risk and places obligations on anyo...
Read definitionAn AI agent is an AI system that autonomously plans and executes multiple steps to reach a goal. It uses a language model as its brain and c...
Read definitionAn AI harness is the software layer around a language model that turns it into a working agent. It manages the loop, tools, context, permiss...
Read definitionAI literacy is the knowledge and judgement people need to use AI responsibly: understanding what a model can do, checking its output, protec...
Read definitionAnomaly detection automatically flags data points, events, or patterns that do not fit normal behaviour. It can catch odd invoices, machine ...
Read definitionAnonymisation makes data no longer reasonably linkable to a person. Pseudonymisation replaces identifiers with codes but keeps a route back ...
Read definitionApache Airflow is an open-source workflow orchestrator for batch-oriented data pipelines. You define workflows as Python code, connect tasks...
Read definitionApache Hudi is an open table format for data lakes that makes Parquet files behave like transactional tables. It is strongest where data cha...
Read definitionApache Iceberg is an open table format for large analytical datasets on object storage. It adds snapshots, schema evolution, partition evolu...
Read definitionApache Kafka is an open-source event streaming platform. Producers write events to topics, Kafka stores them as durable partitioned logs, an...
Read definitionApache Spark is an open-source engine for large-scale data processing. It lets teams write SQL, Python, Scala, Java, or R code while Spark d...
Read definitionAn API, or Application Programming Interface, is a contract that lets software talk to other software. In a data context, APIs are how conne...
Read definitionAn API gateway is a single entry point that sits in front of one or more backend services. It handles cross-cutting work like authentication...
Read definitionAn approval workflow routes a request to the people who must sign off before it takes effect. A purchase, an expense, a data change, or an a...
Read definitionArtificial intelligence is technology that teaches computers to learn, reason, and make decisions from data instead of following hand-writte...
Read definitionAt-least-once delivery means every message should arrive one or more times. It protects against loss, but retries can create duplicates. Tha...
Read definitionAt-most-once delivery means a message is delivered zero or one time. It will not be duplicated, but it can be lost. That trade-off fits tele...
Read definitionAn audit trail is an immutable, time-ordered record of who did what, when, and to which record. You use it to reconstruct events after a dis...
Read definitionBanking as a Service lets non-bank companies offer banking features under their own brand by using a licensed bank or regulated provider beh...
Read definitionA backfill is the rerun of a data pipeline over historical periods that were missed, wrong, or not yet processed. It is common when launchin...
Read definitionBatch processing collects a set of work and runs it as one bounded job on a schedule or a trigger, instead of handling each record the momen...
Read definitionBias in AI is a skew that creeps into models through data, algorithms, or human choices. It is not always harmful, but it has to be managed ...
Read definitionBottleneck analysis finds the step in a process where work gets stuck waiting, the step that dictates total throughput time. You spot bottle...
Read definitionA bridge table is an intermediate table that resolves a many-to-many relationship in a star schema. It stores the valid pairs between two ta...
Read definitionA business glossary is the agreed list of what your business words mean. It records terms like revenue, active customer, margin and churn in...
Read definitionBusiness intelligence (BI) turns data from systems such as ERP, CRM, ecommerce, and accounting into reports and dashboards people use to mak...
Read definitionA business key is the identifier a business already uses to recognise something real, like an invoice number, VAT number, EAN or employee nu...
Read definitionBPMN is a visual standard for mapping business processes clearly. With a fixed set of symbols you can see at a glance who does what, where d...
Read definitionA calculation group applies one DAX pattern to every measure in your model. You write YTD, MTD and YoY% once instead of repeating them for e...
Read definitionA candidate key is any column or set of columns that can uniquely identify every row in a table, with no redundant column in it. A table can...
Read definitionCardinality in Power BI describes how rows in two tables relate: one-to-many, many-to-one, one-to-one, or many-to-many. It controls relation...
Read definitionA case ID is the key that ties all events of one process run together. Think of an order number, a ticket number or a patient file. Without ...
Read definitionChain-of-thought prompting asks a language model to work through intermediate reasoning steps before answering. It can improve arithmetic, l...
Read definitionChange Data Capture (CDC) is the practice of detecting every change in a source system and forwarding it to downstream systems. It keeps you...
Read definitionA chatbot is software that has a conversation with a user through text or speech. Simple bots follow fixed rules and menus; modern bots ofte...
Read definitionCheckpointing periodically saves a streaming job's progress, its input offsets and its running state, to durable storage so it can restart f...
Read definitionChunking is the step where documents are split into smaller pieces before they are embedded and indexed for retrieval. Good chunks help a RA...
Read definitionA circuit breaker watches the calls your application makes to a dependency and, once failures cross a threshold, stops sending calls for a w...
Read definitionClass imbalance is when one class in a classification dataset vastly outnumbers the other, like fraud among normal transactions. The rare cl...
Read definitionA code set is the fixed list of permitted codes and their meanings for a single field, such as order statuses, country codes, or currencies....
Read definitionColumnar storage keeps the values of each column together on disk instead of storing whole rows side by side. An analytical query that reads...
Read definitionA common table expression, or CTE, is a named temporary result set that exists only for the single SQL statement that defines it. You write ...
Read definitionA composite key uses two or more columns together to identify a row. No column is unique on its own, but the combination is. You meet them i...
Read definitionA composite model in Power BI mixes storage modes in one semantic model. Some tables can be imported into memory, while others stay live thr...
Read definitionComputer vision is software that interprets images and video. It can classify images, detect objects, segment defects, read text with OCR, o...
Read definitionConformance checking compares how a process actually runs against how it is supposed to run. It is the second pillar of process mining along...
Read definitionA conformed dimension is a shared dimension table, such as Date, Customer, or Product, reused consistently across several fact tables or dat...
Read definitionA confusion matrix is a small table that splits a classification model's predictions into four counts: right and wrong on each class. It sho...
Read definitionA connection pool is a set of open database connections that an application borrows, uses for one query or transaction, and returns instead ...
Read definitionContext engineering is the practice of deciding what goes into an AI model's context window on each call: the instructions, retrieved docume...
Read definitionThe context window is the amount of text a language model can see and process in a single call. It sets how many instructions, documents, an...
Read definitionA CRM system stores and manages customer relationships: contacts, accounts, leads, opportunities, service cases, campaigns, and customer com...
Read definitionCross-validation estimates how well a model generalises by splitting the data into k folds, training on all but one and scoring on the fold ...
Read definitionA custom connector is one you build yourself when no ready-made connector exists for a source or destination. In Power Automate and Logic Ap...
Read definitionA DAG is a directed acyclic graph: a set of steps connected by one-way dependencies with no loops. In data work, it is the shape used to des...
Read definitionDagster is an open-source data orchestrator built around data assets. You describe the tables, files, models, and checks that should exist, ...
Read definitionA data catalog is a searchable inventory of an organisation's data assets. It records metadata such as owner, schema, description, lineage, ...
Read definitionA data clean room is a controlled environment where two or more parties analyse their combined data without either side seeing the other's r...
Read definitionA data consumer is any person, team, report, dashboard, model, API, or downstream process that reads a data product. Naming your consumers i...
Read definitionA data contract is an explicit agreement between the producer and the consumers of a dataset: which schema, which quality, which frequency, ...
Read definitionA data domain is a clearly bounded business area, such as sales, finance, logistics, or HR, together with the data that belongs to it and th...
Read definitionData downtime is the time during which data is missing, stale, wrong, duplicated, incomplete, or otherwise untrusted. It treats broken data ...
Read definitionData drift is a change in the statistical distribution of the input data a model or report was built on, while the columns and their types s...
Read definitionA data fabric is an architecture that connects distributed data across systems, clouds, and domains through metadata, governance, automation...
Read definitionData freshness measures how recent data is compared with the real-world event or source it represents. A data SLA makes that expectation exp...
Read definitionData governance is the set of roles, rules, processes, and controls that decides how an organisation manages, protects, defines, shares, and...
Read definitionA data lake is a central store for raw data in many formats: tables, JSON, logs, images, documents, audio, and video. It keeps data cheaply ...
Read definitionData leakage is when information that would not be available at prediction time slips into training, so a model scores brilliantly in valida...
Read definitionData lineage shows the full journey data takes inside an organisation. From the original source to the final report, with meaning and contex...
Read definitionData literacy is the ability to read, question, interpret, and communicate with data. It is not the same as being an analyst; it is knowing ...
Read definitionA data mart is a focused analytics layer for one department, domain, or topic. It usually contains a selected subset of warehouse or lakehou...
Read definitionData masking replaces sensitive values with realistic but fake ones so people can work with data without seeing the real content. Static mas...
Read definitionData mesh is an organisational model for data in which each business domain owns its datasets and offers them as products. It breaks with th...
Read definitionData minimisation means collecting and keeping only the personal data you actually need for a specific, stated purpose. It is a core GDPR pr...
Read definitionData observability is continuous monitoring of data health: freshness, volume, schema, distribution, and lineage. It helps teams detect brok...
Read definitionData orchestration schedules, orders, runs, retries, and monitors data tasks as one chain: loads, transformations, quality checks, notebooks...
Read definitionData ownership means that a person or team is formally responsible for a dataset: its definition, quality, access rules, lifecycle, and prio...
Read definitionA data pipeline is an automated chain that moves data from a source to a destination: extracting, transforming, loading, scheduling, logging...
Read definitionA data product is a dataset, semantic model, API, or reporting layer that a team owns and supports for others to use, with a defined consume...
Read definitionData profiling means systematically inspecting a dataset before you build on it: which values occur, how many fields are empty, whether keys...
Read definitionData quality is the degree to which data is fit for the purpose you use it for. Addresses, VAT numbers, customer IDs, dates, and product cod...
Read definitionData reconciliation is the systematic comparison of two datasets or systems to confirm that they match, and to find exactly where they do no...
Read definitionData residency is the physical country, region, or geography where data is stored and processed. In cloud platforms it is usually controlled...
Read definitionData retention is the set of rules that decide how long you keep each type of data and what happens when that time is up: delete, archive, o...
Read definitionData sampling means working with a subset of a dataset instead of all of it, either to move faster and cheaper or to estimate a property of ...
Read definitionA data sharing agreement is the written deal between organisations that share data: what moves, to whom, for what purpose, on what legal bas...
Read definitionData skew is when rows spread unevenly across partitions, so one task in a distributed engine like Apache Spark does most of the work while ...
Read definitionData sovereignty is the question of which laws, authorities, and jurisdictions can affect your data. Physical location matters, but the prov...
Read definitionA data steward looks after the quality and meaning of a specific area of data, such as customer, product, or finance data. In an SME it is u...
Read definitionData testing means running automated checks that verify whether data follows rules you defined in advance: no empty keys, no duplicate IDs, ...
Read definitionData transformation is any operation that changes data on its way from a raw source to a form fit for analysis: casting types, cleaning valu...
Read definitionA data vault is a data warehouse modelling method, introduced by Dan Linstedt, that stores historised and auditable source data in hubs, lin...
Read definitionData virtualization lets you query and combine data from several systems through a logical layer, without first copying all of it into a cen...
Read definitionA data warehouse is a central database that collects data from many source systems and structures it for reporting and analysis. It's optimi...
Read definitionA database index is a separate, sorted structure a database keeps next to a table, so it can find rows without scanning every one. Reads on ...
Read definitionDatabase normalisation organises data into related tables so each fact is stored once. It keeps operational systems consistent, while report...
Read definitionDataflow Gen2 is Microsoft Fabric's self-service ETL and data-preparation item. You transform data with Power Query in the browser and write...
Read definitionDataOps brings software delivery habits into data work: version control, automated tests, separate environments, CI/CD, monitoring, and clos...
Read definitionDAX is the formula language behind Power BI, Excel Power Pivot and Analysis Services. You use it to build calculations like totals, margins ...
Read definitiondbt is a SQL-first transformation tool for analytics teams. You write SELECT statements, and dbt handles dependencies, materialisation, test...
Read definitionA dead-letter queue (DLQ) is a separate queue where a message lands after it fails processing too many times or expires, so one poison messa...
Read definitionA deadlock is a standoff between two database transactions, each holding a lock the other needs, so neither can move. The database breaks th...
Read definitionDebezium is an open-source platform for log-based change data capture. It reads database transaction logs and emits inserts, updates, and de...
Read definitionDeduplication finds records that refer to the same customer, product, supplier, or other entity, then merges or removes the duplicates so on...
Read definitionDeep learning is machine learning with neural networks that have many layers. Those layers learn increasingly abstract patterns in images, s...
Read definitionA degenerate dimension is a dimension key that stays in the fact table without its own dimension table, such as an order number, invoice num...
Read definitionDelta Lake is an open storage format that extends plain Parquet files with transactions, schema enforcement, and time travel. It forms the f...
Read definitionDenormalisation deliberately stores data redundantly, in wider tables and repeated values, so reports read faster with fewer joins. You acce...
Read definitionA deployment pipeline is the built-in ALM tool of Microsoft Fabric. You build in a Development workspace, test in a Test workspace, and roll...
Read definitionA derivation rule defines how a field is calculated from other data instead of stored or copied directly. It pins down the inputs, the formu...
Read definitionDifferential privacy is a mathematical privacy guarantee: an analysis should give almost the same result whether one person's data is includ...
Read definitionA digital twin is a digital copy of a physical object, process, or system, fed by live data. You use it to run simulations, test scenarios, ...
Read definitionA dimension table holds the descriptive attributes, like customer, product, date, region or channel, that give context to the numbers in a f...
Read definitionDimensional modelling is Ralph Kimball's design method for analytics: fact tables for events, dimension tables for context, usually arranged...
Read definitionDirect Lake is a storage mode for Power BI that reads straight from Delta tables in OneLake. You get the speed of Import without the refresh...
Read definitionDirectQuery is a storage mode in Power BI and Fabric that keeps no copy of the data and instead sends every query straight to the source. Yo...
Read definitionDisaster recovery is the plan to bring systems and data back after a major failure. RPO sets how much data you can afford to lose, counted b...
Read definitionA DPIA is a structured risk analysis you run on a data project before you build it, when the processing is likely to create a high risk to p...
Read definitionEmbeddings turn words, sentences, or images into numbers that capture their meaning. Neural networks learn them from huge amounts of text. T...
Read definitionEntity resolution decides which records refer to the same real-world person, company, product, or asset across different systems. It links J...
Read definitionAn ERP system records a company's core processes: sales, purchasing, stock, production, projects, invoicing, and accounting. For reporting, ...
Read definitionETL and ELT stand for Extract, Transform, Load and Extract, Load, Transform. They are two ways of moving data from source systems into a cen...
Read definitionEvals are structured tests for measuring the quality of a language model or AI application. They give you a repeatable way to compare prompt...
Read definitionAn event log is the foundation of process mining. It records every step that happens in a process, showing what was done, when, and to which...
Read definitionEvent sourcing stores every change to your data as an ordered sequence of events and treats that log as the source of truth. Instead of keep...
Read definitionIn event-driven architecture, services talk by emitting events like order.placed instead of calling each other directly, and anything intere...
Read definitionAn Eventstream is the Microsoft Fabric item that captures real-time events, applies light transformations, and routes them to destinations s...
Read definitionExactly-once processing means every event has exactly one effect in the result, even if messages are retried after failures. It is less abou...
Read definitionAn execution plan is the tree of operators a database runs to answer a query: the scans, joins, sorts, and aggregates the query optimizer ch...
Read definitionExperiment tracking records every model training run with its parameters, code version, dataset version, metrics, and output files. It lets ...
Read definitionExponential backoff is a retry strategy where you double the wait after each failed attempt, up to a maximum delay, so a struggling service ...
Read definitionA Fabric capacity is the compute pool that Power BI, lakehouses, notebooks and pipelines run on in Microsoft Fabric. The F-SKUs run from F2 ...
Read definitionA fact table stores the numbers a business measures, such as orders, invoice lines, payments, or stock counts, with one row per event and fo...
Read definitionA factless fact table is a fact table without numeric measures. It stores only keys to dimensions, recording that an event happened or that ...
Read definitionFeature engineering turns raw columns into the inputs a model can learn from. From an orders table you build a feature like the number of or...
Read definitionA feature store is a governed place to define, store, share, and serve machine-learning features. Its job is to keep training and inference ...
Read definitionFew-shot prompting means showing a language model a few worked examples before asking the real question. Instead of explaining every rule, y...
Read definitionFine-tuning is the practice of further training an existing AI model on your own data so it handles a specific task, tone, or domain better....
Read definitionA foreign key is a column whose values must match an existing row in another table, usually that table's primary key. It is how a relational...
Read definitionGDPR is the European regulation that governs how organisations process personal data. It does not ban data use, but it requires a clear purp...
Read definitionGenerative AI (GenAI) is technology that produces new content on its own, things like text, images, code, or music. It learns patterns from ...
Read definitionGit is a distributed version control system that records changes to files over time. Teams use it to track who changed what, review work bef...
Read definitionA golden record is the resolved, best-value version of one business entity, such as a customer or product, assembled from many source system...
Read definitionA GPU is a powerful chip built for fast, parallel calculations. It runs thousands of small operations at the same time, which makes it ideal...
Read definitionGranularity, or grain, is the level of detail in a table: what does one row represent? Declaring the grain first prevents double counting, m...
Read definitionGrounding is the property of an AI answer being tied to source material you supply, a document, a database record, a tool result, so each cl...
Read definitionAI guardrails are the protective layers around an AI application: input checks, system prompts, output filters, moderation services, tool pe...
Read definitionA hallucination is when an AI model says something that sounds plausible but isn't true. The model invents details that aren't in any source...
Read definitionThe happy path is the ideal route through a process or system, without exceptions, errors or detours. It describes what happens when everyth...
Read definitionHeadless BI separates the layer that defines business metrics from the tool that visualises them. Revenue, active customers, or churn are de...
Read definitionHuman-in-the-loop (HITL) means that people stay actively involved in decisions and processes that are partly handled by AI. They review, cor...
Read definitionHybrid search combines keyword search, often BM25, with vector search over embeddings. It returns exact matches and semantic matches in one ...
Read definitionAn operation is idempotent when running it once or many times leaves the system in the same final state. Idempotence makes retries safe afte...
Read definitionImport mode is Power BI's default storage mode. On refresh, the semantic model copies data from the source, compresses it with VertiPaq, and...
Read definitionAn incremental load moves only the rows that changed since the last successful run, instead of rereading the whole source. It keeps pipeline...
Read definitionIncremental refresh is the Power BI and Fabric feature that processes only new or changed data at refresh time instead of reloading the whol...
Read definitionInference is the moment a trained AI model is used: new input goes in, a prediction, classification, image, or answer comes out. Training is...
Read definitionIntelligent Process Automation (IPA) combines automation with artificial intelligence to make business processes smarter and more efficient....
Read definitionAn iPaaS is a cloud platform for connecting business applications through managed connectors, workflows, transformations, scheduling, monito...
Read definitionJevons paradox is the idea that when something becomes more efficient, we often end up using more of it instead of less.
Read definitionJSON is a text format for exchanging structured data between systems. It is readable enough for humans, strict enough for machines, and show...
Read definitionA junk dimension is a small dimension table that groups loose low-cardinality flags, indicators, and status codes. It keeps the fact table n...
Read definitionA knowledge graph stores information as entities and named relationships instead of only rows and columns. It is useful when the question is...
Read definitionA KPI is a measurable indicator of progress toward an important goal. It is not just any metric. A good KPI has a clear definition, owner, t...
Read definitionKQL, short for Kusto Query Language, is Microsoft's query language for time-series and log data. You meet KQL inside Azure Data Explorer, Fa...
Read definitionA lakehouse combines the flexibility of a data lake with the structure and performance of a data warehouse. You store raw and processed data...
Read definitionA Large Language Model (LLM) is an AI model trained on enormous amounts of text that understands and generates language. LLMs are the engine...
Read definitionLate arriving dimensions occur when a fact reaches the warehouse before the matching dimension row exists. Kimball's standard fix is an infe...
Read definitionLeast privilege means every identity, human or machine, gets the minimum access it needs to do its job, and no more. It limits the damage wh...
Read definitionLow-code and no-code tools let people build apps, automations, dashboards, and data transformations with visual building blocks instead of t...
Read definitionMachine learning is software that learns patterns from examples instead of following hand-written rules. You train a model on historical dat...
Read definitionA mapping rule is the part of a source-to-target mapping that says how a value gets from a source field to a target field: a direct copy, a ...
Read definitionMaster Data Management (MDM) is the discipline of keeping one trusted version of core business data, such as customers, products, suppliers,...
Read definitionA materialized view is a query whose result is stored as a table, so reads are fast, at the cost of freshness: the data is only as current a...
Read definitionMCP, short for Model Context Protocol, is an open standard from Anthropic that lets AI models talk to tools and data in a uniform way. One p...
Read definitionMedallion architecture organises a lakehouse into three layers: bronze for raw data, silver for cleaned data, and gold for business-ready ta...
Read definitionA message queue sits between two systems: a producer drops a message in, and a consumer picks it up when it is ready. That decouples the two...
Read definitionMetadata is data about data: the information that explains what a dataset is, where it came from, how fresh it is, who owns it, and how it m...
Read definitionA metrics layer is the central place where business metrics such as revenue, active customers, margin, and churn are defined once, then reus...
Read definitionMicrosoft Copilot is the family name for every AI assistant Microsoft builds into its products, from Word and Excel through to Power BI and ...
Read definitionMicrosoft Fabric is an all-in-one cloud platform that brings data storage, transformation, and reporting together in one place. It connects ...
Read definitionMicrosoft Power Platform is Microsoft's family of low-code products for building apps, automating workflows, analyzing data, publishing webs...
Read definitionMicrosoft Purview is Microsoft's platform for data governance, data security, and compliance. It helps organisations see what data they have...
Read definitionMLOps is the way of working that brings machine learning models reliably into production and keeps them healthy there. It combines elements ...
Read definitionModel distillation trains a smaller student model to imitate a larger teacher model. The student learns from the teacher's outputs, so it ca...
Read definitionModel drift happens when an AI model gets worse at predicting because data or real-world patterns shift over time. Regular retraining and co...
Read definitionA model registry is a versioned catalogue of trained machine learning models. Each version records which data, code and run produced it, so ...
Read definitionModel serving is how a trained machine learning model is made available to return predictions, either as a real-time endpoint that answers e...
Read definitionA multimodal model is an AI model that can work with more than one type of input, such as text, images, audio, and video. Some models can al...
Read definitionNatural Language Processing (NLP) is the umbrella name for techniques that let computers understand human language. You use it every day wit...
Read definitionA neural network is a computer model that learns by working through examples, loosely inspired by the human brain. It is built from layers o...
Read definitionNL2SQL turns a plain-language question into a SQL query against a known schema. The demo looks like magic, but getting the joins, column mea...
Read definitionA data notebook is an interactive work document with code, explanation, and results in cells. Analysts and data scientists use notebooks to ...
Read definitionOAuth is the protocol behind buttons like "Sign in with Google" and "Allow access to your Drive". It lets an app get limited, revocable acce...
Read definitionObject Level Security hides entire tables or columns in your data model from specific users. When sensitive fields like salaries, margins or...
Read definitionObject-centric process mining is an approach where events are not tied to a single case but to several business objects at once: an order, t...
Read definitionOCR, short for Optical Character Recognition, is technology that reads text inside an image or scanned document. It turns letters and number...
Read definitionOLTP systems run day-to-day transactions such as orders, payments, and stock changes. OLAP systems analyse that data over time. The differen...
Read definitionAn on-premises data gateway is a Windows service that connects Microsoft cloud services such as Power BI, Fabric, Power Automate, and Power ...
Read definitionOneLake is Microsoft’s central data storage system within Microsoft Fabric. It serves as the single, unified location where all business dat...
Read definitionA OneLake shortcut is a pointer from your Fabric lakehouse or KQL database to data that lives somewhere else, without copying it. Internal O...
Read definitionAn ontology is a formal shared model of the concepts in a domain: what kinds of things exist, which properties they have, and how they relat...
Read definitionAn open-weight model is an AI model whose trained weights can be downloaded and run outside the vendor's hosted API. The licence still decid...
Read definitionA paginated report is a pixel-precise report designed for printing, PDF export, or formal distribution. It has fixed pages, headers, footers...
Read definitionpandas is the Python library many analysts use for tabular data work. Its Series and DataFrame objects let you read, clean, join, reshape, a...
Read definitionApache Parquet is a columnar file format designed for analytical queries on large datasets. It is the storage layer under almost every moder...
Read definitionPartitioning splits a large table into smaller physical parts based on a column such as date, month, region, or tenant. When queries filter ...
Read definitionPII, or personally identifiable information, is any data that can single out a specific person, directly or indirectly. In Europe the term t...
Read definitionPower Automate is a Microsoft tool that lets you automate repetitive tasks and processes without writing code. You build flows visually that...
Read definitionPower BI is Microsoft’s business intelligence tool that helps people understand their data through interactive dashboards and reports. It co...
Read definitionPower BI Embedded lets you build Power BI reports straight into your own software or website, so users see insights without ever opening Pow...
Read definitionPBIR is the new default file format for Power BI reports. Instead of storing everything in one large JSON file, a report is now split into m...
Read definitionPBIP is the folder-based project format for Power BI. Instead of storing everything in one PBIX file, a PBIP project splits the semantic mod...
Read definitionPower Query is Microsoft's transformation language for data. You use it visually in Power BI, Excel, and Fabric Dataflows, with M as the und...
Read definitionPrecision and recall are two scores that grade a classification model. Precision is the share of flagged cases that were right, TP / (TP + F...
Read definitionPredicate pushdown pushes a query's filter down to the data source or file reader, so rows that fail the condition are never read. In Parque...
Read definitionPredictive analytics uses historical data to estimate what is likely to happen next: future sales, churn risk, late payment, stock needs, or...
Read definitionA primary key is the column, or set of columns, that uniquely identifies every row in a table. The database enforces two rules on it: the va...
Read definitionProcess automation is the use of software to take over repetitive tasks that people would normally do by hand. It connects systems, moves da...
Read definitionProcess discovery is the step in process mining that automatically draws a process map from your event log. No workshops or interviews, just...
Read definitionProcess mining helps companies see how their processes truly run by using real data from their systems. It turns digital traces from tools l...
Read definitionPrompt engineering is the craft of writing instructions for an AI model so it produces useful output. A good prompt largely decides whether ...
Read definitionPrompt injection is an attack where a user or external content tricks a language model into ignoring, overriding, or leaking its instruction...
Read definitionA prompt template is a reusable prompt with named placeholders that get filled at runtime, so one set of instructions runs over many inputs....
Read definitionQuantisation makes an AI model smaller by storing its weights with fewer bits, often 8-bit or 4-bit instead of 16-bit. It reduces memory use...
Read definitionQuery folding is what lets Power Query turn your transformation steps into a single query that the data source runs itself, so filters, colu...
Read definitionA query optimizer is the part of a database that decides how a SQL query actually runs: which indexes to use, how to join the tables, and in...
Read definitionRAG or Retrieval-Augmented Generation is a technique where an AI model first looks up relevant information in a dedicated knowledge base bef...
Read definitionA rate limit is a cap on how many requests a client can send to an API within a set time window, such as 5,000 per hour or 100 per second. P...
Read definitionRBAC attaches permissions to roles and gives people roles, so access follows job function instead of a list kept per person. Add someone to ...
Read definitionA read replica is a read-only copy of a database that serves reports, exports and dashboards so the primary can focus on writes. Replication...
Read definitionReference data is the set of shared, mostly standardised lookup values that give other data meaning: country codes, currencies, order status...
Read definitionReinforcement learning is a form of machine learning where systems learn by experience. They try things, get feedback, and gradually get bet...
Read definitionReranking is a second scoring step in search and RAG. A fast search retrieves a broad candidate list, then a slower reranker compares each c...
Read definitionA retry policy is the set of rules a system follows when an operation fails: which errors to retry, how many attempts, how long to wait, and...
Read definitionReverse ETL syncs modelled data from a warehouse or lakehouse back into operational tools such as CRM, marketing automation, support, and ad...
Read definitionA rework loop is a pattern where the same activity runs more than once within a single case: an invoice that gets re-created, a file that go...
Read definitionA ROC curve plots a classification model's true positive rate against its false positive rate at every threshold, so it measures ranking qua...
Read definitionA role-playing dimension is one dimension table used several times by the same fact table, each time in a different role. The classic exampl...
Read definitionRow Level Security makes sure users only see the rows of your dataset they are allowed to see. It is a clean way to share one dataset safely...
Read definitionRPA, or Robotic Process Automation, is a technology that uses software bots to perform repetitive computer tasks automatically. It helps bus...
Read definitionA runbook is a written procedure for a recurring operational task or a known failure, written so that someone who did not build the system c...
Read definitionThe saga pattern runs a business process that spans several services as a sequence of local transactions, each paired with a compensating ac...
Read definitionSchema drift happens when the structure of a source changes over time: columns are added, removed, renamed, reordered, or change type. Pipel...
Read definitionSchema evolution is how you change the shape of a table or message stream over time, adding a column, renaming one, or changing a type, with...
Read definitionA schema registry is a central store for the message schemas on a streaming platform. It versions every schema and checks each new one again...
Read definitionSecret management is how you store and control access to credentials, API keys, connection strings, and certificates outside your code, then...
Read definitionSelf-service analytics gives people in the business the tools to work with data themselves, without depending on IT for every question. With...
Read definitionA semantic model is a layer on top of your data that translates tables, relationships, and calculations into business language. It's the bri...
Read definitionA service level objective, or SLO, is an internal target for how reliable a service has to be, written as a measurable number over a time wi...
Read definitionA single source of truth (SSOT) is the agreement that every fact in your organisation has one authoritative place everyone reads from, so re...
Read definitionSlowly Changing Dimensions (SCD) is the set of patterns for storing changes in dimension data (customers move, products change category) cor...
Read definitionThe small files problem is when a data lake or lakehouse table ends up spread across thousands of tiny files. The query engine then spends m...
Read definitionA small language model (SLM) is a language model compact enough to run on modest hardware you control: a single GPU, a laptop, sometimes a p...
Read definitionA snapshot fact table records the state of something at a fixed moment instead of the individual events behind it. It is one of Kimball's th...
Read definitionSnowflake is a modern cloud-based data warehouse that helps organisations store and analyse large volumes of data efficiently. It separates ...
Read definitionA snowflake schema is a dimensional model where dimension tables are normalised into linked sub-dimensions, so a single entity like product ...
Read definitionA soft delete marks a row as deleted with a flag or timestamp instead of removing it, so the record stays in the database for recovery and h...
Read definitionA source field is a field exactly as it exists in the system that produced it, with its own name, data type, length and encoding, before any...
Read definitionSpeech-to-text (STT) turns spoken words into readable text. You can use it to convert calls, recordings, or meetings into notes and reports ...
Read definitionSQL is the standard language for querying and managing relational databases. You describe the result you want, and the database engine works...
Read definitionA star schema is a data model where one central fact table is surrounded by dimension tables. It is the standard pattern for BI and delivers...
Read definitionA stored procedure is a named routine of SQL statements saved inside the database. You call it by name, pass parameters, and the database ru...
Read definitionStreaming data is data processed continuously as each event arrives, not gathered into scheduled batches. It brings its own concepts: event ...
Read definitionStructured output means making a model return machine-readable data that fits a schema, usually JSON, instead of free text. Modern APIs can ...
Read definitionSupervised learning is a way for computers to learn from examples that already include the right answer. The model studies thousands of labe...
Read definitionA surrogate key is an artificial, meaningless integer that your warehouse assigns to every row in a dimension table. It becomes the dimensio...
Read definitionSynthetic data is artificially generated records that copy the statistical shape of a real dataset without being real people. You use it for...
Read definitionA system of record is the application your organisation treats as authoritative for a specific kind of data, such as the ERP for invoices or...
Read definitionA system prompt is the instruction an application gives a language model before the conversation starts, setting its role, scope, tone, outp...
Read definitionTable compaction rewrites the many small files a lakehouse table piles up into fewer, larger ones, so scans stop paying per-file overhead. D...
Read definitionA target field is the column in a destination schema that receives a value after a source field is mapped, transformed, or derived. Unlike t...
Read definitionTask mining captures how work happens on the desktop: mouse clicks, keystrokes, window switches and application use at the employee's own ma...
Read definitionTemperature is the sampling setting that controls how random a large language model's output is. It scales the token scores before the softm...
Read definitionText-to-speech (TTS) turns written text into spoken audio. It evolved from clunky mechanical machines to modern AI voices that sound almost ...
Read definitionThroughput time is the time between the start event and the end event of a case: the full clock on an order, request or ticket, including ev...
Read definitionTime travel lets you query a table as it looked at an earlier version or timestamp. Delta Lake and Apache Iceberg keep the older versions th...
Read definitionA token is the smallest unit of text an AI model processes, usually half a word or a punctuation mark. Tokens drive both the bill on every A...
Read definitionTool use, which most model vendors call function calling, lets a language model do more than write text. You describe a set of tools and the...
Read definitionA train-test split holds part of your data back from training so a model is scored on examples it never saw. That test score estimates how i...
Read definitionTransaction isolation decides how much one running database transaction can see of others that are still in flight. It is the I in ACID, and...
Read definitionA transactional outbox writes an outgoing event into a table in the same database transaction as the business change, so the two commit toge...
Read definitionThe transformer architecture is the engine behind modern AI. It processes text, images and sound by understanding the relationships between ...
Read definitionUnsupervised learning is a form of machine learning where computers find patterns in data on their own, without anyone telling them what is ...
Read definitionAn upsert is a single database write that inserts a row when it does not exist and updates it when it does, matched on a key you name. Data ...
Read definitionA validation rule is a test that judges a value or a row and returns pass or fail, without changing the data. It turns data quality into som...
Read definitionVariant analysis shows how many different routes the same process actually takes. You see which sequence of steps is most common, how long e...
Read definitionA vector database stores embeddings and searches them at high speed on meaning rather than exact words. They form the memory layer under RAG...
Read definitionVector search finds the items whose vectors sit closest to a query vector, ranking results by meaning instead of exact words. Exact search c...
Read definitionVertiPaq is the in-memory columnar engine behind Power BI import mode, Analysis Services Tabular, and Excel Power Pivot. It copies your data...
Read definitionIn stream processing, a watermark is the engine's assertion about how far event time has advanced, used to decide when a time window can clo...
Read definitionA webhook is an HTTP callback a provider sends to your URL the moment an event happens, so you stop polling for changes. The receiver's job ...
Read definitionA window function computes a value across a set of related rows, such as a running total or a rank, without collapsing them the way GROUP BY...
Read definitionA workflow engine is software that runs business processes step by step in a reliable way. It tracks where each process is, decides what sho...
Read definitionZ-ordering rearranges a Delta Lake table so that rows you often query together share the same Parquet files. That lets data skipping drop fi...
Read definitionZero-trust data architecture is a security model that trusts nobody by default. Every user and every application has to prove themselves aga...
Read definition