Insights
Data architecture, Tableau, cloud engineering, and BI — from engineers with 10+ years of enterprise experience.
A synthesis of Bain and McKinsey research on what is actually stopping enterprise AI from delivering. Eight structural problems — and what data leaders are doing about each one.
The CFO is the most underused ally in enterprise AI. They control the most governed data in the business, sign off on every system that creates a data footprint, and carry the commercial authority to make data standards stick. Here is how to make the case.
Agentic AI does not make requests — it takes actions autonomously across systems. The architecture handling your dashboard queries was never designed for that. Here is what needs to change.
After years of managing both environments at enterprise scale, here is our honest assessment of when to migrate to Tableau Cloud — and when to stay on Server.
Day rates, project fees, retainer structures — a transparent breakdown of what senior Tableau consulting actually costs and what drives the price.
Most data cost problems are architecture problems in disguise. Here are the seven clearest signals that your data infrastructure needs a structural rethink.
Slow Tableau Server performance is almost always fixable. This is the diagnostic framework we use to identify and resolve the most common bottlenecks.
What we have learned from building production Azure data platforms at enterprise scale — the patterns that work and the ones that cost you later.
These terms are often used interchangeably but they describe different disciplines. Understanding the difference matters when you are building your data team.
What a data architect actually does, the signals that separate strong candidates from plausible ones, what to pay, and whether to hire in-house, engage a contractor, or work with a consulting firm.
A transparent breakdown of what data architecture consulting actually costs — from $15k assessments to $500k platform builds. What drives the price, what red flags look like, and how to get a fair proposal.
Salesforce has announced the end of life for Tableau Server. Here is the EOL timeline, your three options, what a migration actually involves week by week, the common blockers, and what to do this week.
We work with both platforms every day. Here is a direct, experience-based comparison — where each platform wins, where it loses, what the real migration costs look like, and how to make the decision without getting sold to.
The data lakehouse pattern combines the storage economics of a data lake with the query performance and governance of a data warehouse. Here is what the architecture actually looks like, when it is the right choice, and what it takes to build one at enterprise scale.
Data governance is not a compliance project. It is the set of policies, ownership structures, and technical controls that make your data trustworthy enough to act on. Here is what it actually involves — and what most implementations get wrong.
We build on both platforms every week. Here is a direct, experience-based comparison — what each is genuinely better at, the common misconceptions, pricing reality, and how to make the decision for your specific workloads.
A semantic layer sits between your data platform and your BI tools, translating raw tables into business-ready metrics with consistent definitions. Here is what it is, what it does, how to build one, and why most enterprise data quality problems are semantic layer problems in disguise.
ETL and ELT describe two different approaches to moving and transforming data. The right choice depends on your data volumes, transformation complexity, and cloud platform. Here is a practical breakdown of when each pattern fits and what drives the decision.
A data warehouse is a central repository for structured, integrated data built for analytical querying. Here is how modern cloud data warehouses work, how they differ from data lakes and lakehouses, and the decision framework for which architecture fits your organisation.
Microsoft Fabric consolidates Power BI, Azure Synapse, Azure Data Factory, and other Microsoft data services into a single platform. Here is what it actually includes, what it is genuinely better at, and the honest assessment of when migration makes sense — and when it does not.
Data mesh is an organisational and architectural approach that distributes data ownership to domain teams instead of centralising it in a data engineering function. Here is what it actually involves, who it is designed for, and the honest assessment of when it solves a real problem versus adding complexity.
Four platforms dominate enterprise BI. Each has genuine strengths and real weaknesses that vendor materials will not tell you. Here is an honest, experience-based comparison — what each platform is actually best at, the commercial realities of each, and the decision framework for your organisation.
Master data management (MDM) creates a single authoritative record for core business entities — Customer, Product, Supplier, Location — across all systems. Here is what it involves, what problems it solves, and what implementation actually looks like.
Most Tableau dashboards are built to answer a question. The best ones are built to support a decision. Here are the design principles, layout patterns, and performance considerations that senior Tableau developers use to produce dashboards that executives actually use.
Moving your data infrastructure to the cloud is a multi-month programme that most organisations underestimate. Here is how to plan it, what it costs, the phases that cannot be skipped, and the mistakes that push timelines from 6 months to 18.
Excel is not a BI tool — but it handles a lot of BI work at most organisations. Here is the honest assessment of when Excel is the right answer, when Power BI is genuinely better, and how to make the transition without losing the analytical capability your team already has.
Financial services organisations face data architecture requirements that most enterprise platforms are not designed for: regulatory data lineage, real-time risk, strict access controls, and the need to reconcile trading, risk, and finance data across systems that were never designed to talk to each other.
Most Tableau Server to Cloud migration quotes range from $15,000 to $80,000+. The difference is driven by environment complexity, content volume, authentication requirements, and whether embedded analytics need to be rebuilt. Here is a transparent breakdown of what you are actually paying for.
The modern data stack — Fivetran, Snowflake, dbt, and a BI layer — has become the default architecture for mid-market analytics. Here is what it actually is, how the layers fit together, what it costs to build, and when it is not the right answer.
dbt (data build tool) is the standard transformation layer in modern data stacks. You write SQL SELECT statements; dbt handles execution, testing, documentation, and lineage. Here is what it does, how it works, and where it fits in your data architecture.
Azure Synapse and Databricks are the two dominant enterprise data platforms on Azure. Synapse is SQL-first and Azure-native; Databricks is Spark-native and ML-focused. Here is the honest comparison — and why Microsoft Fabric is changing the decision for new builds.
Kimball and Inmon represent two foundational philosophies for data warehouse design. Kimball is bottom-up, dimensional, and fast-to-value. Inmon is top-down, normalised, and enterprise-integrated. Here is how to choose — and why the modern data stack has changed the decision.
Tableau Prep is Tableau's data preparation tool — designed for analysts who need to clean, shape, and combine data before visualising it. Here is what it does well, where it falls short, and how it fits in a broader data architecture.
Snowflake pricing is consumption-based and can scale unexpectedly if not actively managed. Credits, virtual warehouses, storage, and data transfer all contribute to the bill. Here is how the pricing model works and the specific controls that keep costs predictable.
Power Platform is Microsoft's suite of low-code tools: Power BI (analytics), Power Apps (app building), Power Automate (workflow automation), and Power Pages (web portals). Here is how they relate, when each is appropriate, and how Power BI fits within the broader Microsoft platform.
Tableau Server licensing is role-based and can be complex to optimise as your user base grows. Here is how the licensing model works, what each role includes, and the common mistakes that lead to overpaying or under-licensing.
Data engineers build the pipelines and platforms that make data available. Data scientists build models and analysis on top of that data. The distinction matters for hiring, team design, and understanding why ML projects fail without the right infrastructure beneath them.
BigQuery and Snowflake are the two most widely deployed cloud data warehouses. Both are mature, performant, and well-supported. The decision comes down to your cloud platform, pricing tolerance, SQL dialect requirements, and ML/AI ambitions.
The Tableau REST API lets you automate virtually every administrative task: publishing workbooks, managing users, triggering extract refreshes, querying site content, and more. Here is what the API can do and the patterns that save the most engineering time.
Looker and Tableau represent fundamentally different philosophies about enterprise BI. Looker is code-first and semantic-layer-led; Tableau is visual-first and analyst-led. Here is an experience-based comparison of both platforms in production enterprise environments.
Cloud data infrastructure costs grow faster than most organisations expect. Compute waste, unmanaged storage, over-provisioned warehouses, and unoptimised queries are the primary culprits. Here is the diagnostic framework and the specific controls that reduce costs.
Snowflake's multi-cluster shared data architecture — separated compute and storage, virtual warehouses, micro-partitions — is different from traditional data warehouses in ways that directly affect how you should design schemas, queries, and pipelines for it.
Tableau Embedding API v3 lets you embed interactive Tableau dashboards in any web application. Here is how embedding works, what authentication options are available, and the licensing requirements you need to understand before building an embedded analytics product.
Most data governance programmes fail not because the framework is wrong but because it is not implemented. Here is a governance framework designed for practical implementation: ownership, definitions, quality standards, and access control — with the organisational structures that make them stick.
Azure Data Factory is Microsoft's cloud ETL/ELT and data integration service. It connects 90+ data sources, orchestrates pipeline runs, and integrates with the full Azure data stack. Here is how it works, when it is the right tool, and the patterns that make ADF pipelines maintainable.
Databricks pricing is based on Databricks Units (DBUs) — consumption-based compute billing that varies by cluster type and cloud. Understanding the pricing model and applying the right cost controls is essential before costs scale with your data workloads.
Delta Lake is an open-source storage layer that adds ACID transactions, schema enforcement, and time travel to data lakes built on object storage. It is the foundation of the Databricks Lakehouse and is now supported natively by Snowflake, BigQuery, and other platforms.
Administering Tableau Server well — maintaining performance, managing users, governing content, monitoring health, handling upgrades — is a full-time responsibility. Here is the complete reference for Tableau Server administrators.
Data pipelines that fail silently, produce wrong data, or require constant manual intervention are the primary reason data teams lose stakeholder trust. Here is the engineering discipline that makes pipelines reliable.
Deploying Power BI to enterprise scale — Premium capacity, deployment pipelines, workspace governance, row-level security, gateway management — requires more than publishing reports. Here is the complete deployment framework.
Fivetran and Airbyte are the two dominant data ingestion platforms for loading SaaS and database sources into cloud data warehouses. Fivetran is fully managed with a premium; Airbyte is open-source with operational overhead. Here is the honest comparison.
Apache Airflow is the most widely deployed workflow orchestration platform for data engineering. It schedules, monitors, and manages complex data pipeline dependencies via Python-defined DAGs. Here is how it works and the patterns that make Airflow deployments maintainable.
Snowflake and Amazon Redshift are the two most widely deployed cloud data warehouses. Here is a direct comparison on performance, cost, architecture, and when each platform is the right choice.
The patterns that separate maintainable dbt projects from ones that become technical debt: project structure, naming conventions, testing strategy, documentation, and performance.
The data lakehouse pattern is reshaping how organisations think about analytics infrastructure. Here is when a lakehouse is the right choice versus a traditional cloud data warehouse.
Snowflake, BigQuery, and Redshift bills surprise organisations that build for scale without building for cost. Here is the systematic approach to cutting warehouse spend by 30–50% without degrading analytics performance.
Tableau Server end-of-life makes Tableau Cloud migration a near-term reality for most organisations. Here is the technical migration path, the feature gaps to plan for, and the cost and timeline to expect.
A data catalog is the foundation of enterprise data governance — but the tool market is crowded and the products diverge significantly. Here is a direct comparison of the leading options.
Looker and Power BI are both enterprise BI platforms but with fundamentally different philosophies — Looker is semantic-layer-first, Power BI is self-service-first. Here is a direct comparison for buyers evaluating both.
Understanding how Tableau connects to data — live connections vs extracts, published data sources vs embedded connections, and Tableau Bridge for private networks — is essential for building performant, maintainable analytics.
Data contracts define the agreement between data producers and consumers — schema, SLAs, quality guarantees, and ownership. Here is the practical guide to implementing them in a production data platform.
Microsoft Fabric is the unified analytics platform that replaces Azure Synapse, Azure Data Factory, and Power BI Premium. Here is what Fabric actually includes, what it costs, and whether it is the right choice for your organisation.
Tableau Dashboard Extensions allow you to embed custom web applications, third-party visualisations, and interactive controls directly in Tableau dashboards. Here is how they work and when to use them.
dbt and Spark both transform data, but they serve different use cases. dbt is SQL-first, warehouse-native, and built for analysts. Spark is code-first, distributed, and built for large-scale data engineering. Here is when each is the right choice.
Tableau subscriptions deliver dashboard and view snapshots via email or Slack on a schedule. Here is how to configure them, what the common failure modes are, and how to govern them at enterprise scale.
Choosing a BI tool is a multi-year commitment that affects every analyst in your organisation. Here is the structured evaluation framework — criteria, process, and the questions that reveal real capability.
Data models that work in the prototype almost never survive production unchanged. Here are the principles that produce data models analysts can trust, engineers can maintain, and the business can evolve.
Tableau Sets define a custom subset of dimension members — IN or OUT of a specified condition. Combined with set actions, they enable cohort analysis, top-N comparisons, and highlight interactions that filters cannot produce.
The medallion architecture organises data lakehouse layers by quality and transformation stage — raw in bronze, cleaned in silver, business-ready in gold. This guide covers when to use it, how to implement it, and the common mistakes that break the pattern.
Looker Studio is free and fast to start with. Tableau is expensive and takes longer to deploy. The difference in what you get for that investment is significant — this guide covers where each tool excels and when the cost gap is justified.
Apache Iceberg is the open table format that enables ACID transactions, schema evolution, time travel, and hidden partitioning on object storage. This guide covers what Iceberg actually does, how it compares to Delta Lake and Hudi, and when to use it.
A semantic layer sits between the data warehouse and BI tools, centralising business metric definitions so every tool reports the same numbers. This guide compares the leading semantic layer tools and the trade-offs between them.
Data architects work with a specific set of tools — modeling, documentation, governance, lineage, and infrastructure provisioning. This guide covers the tools in each category, what they are used for, and how mature teams combine them.
DataOps applies DevOps practices — version control, CI/CD, automated testing, observability — to data pipelines. This guide covers what DataOps means in practice, the tools that implement it, and why it matters for data team reliability.
BigQuery charges by bytes scanned. Every architectural decision — how tables are partitioned, whether clustering is applied, how queries are structured — directly affects the bill. This guide covers the key BigQuery design decisions and how to control costs without sacrificing query performance.
A dashboard designed to tell a specific story is very different from one designed for open-ended exploration. Both have their place — but conflating them produces dashboards that do neither well. This guide covers the design principles for each mode.
Redshift performance and cost depend heavily on distribution key and sort key choices made at table creation. This guide covers the distribution styles, sort key types, Redshift Spectrum for external tables, and the common design mistakes that cause performance degradation.
Tableau has extensive mapping capabilities — from basic choropleth maps to custom spatial files, density maps, and dual-layer maps. This guide covers every map type in Tableau, when to use each, and the common configuration errors that produce incorrect geographic results.
AI systems consuming enterprise data create governance requirements that traditional data governance frameworks did not anticipate — training data quality, model lineage, feature store governance, and the audit requirements for AI-driven decisions. This guide covers what changes.
dbt macros are reusable SQL snippets defined with Jinja templating. They eliminate repeated logic in your models, allow dynamic SQL generation, and are the foundation of dbt packages. This guide covers when and how to write macros.
Tableau extracts are local copies of data stored in the Hyper columnar format. They are the primary performance lever for Tableau dashboards. This guide covers extract creation, size optimisation, incremental refresh, and when to use live connections instead.
Data mesh is a compelling architectural pattern but notoriously difficult to implement. This guide covers the practical steps: defining data domains, establishing product ownership, building the self-serve platform, and the federated governance model that makes it work.
Snowflake costs can scale unexpectedly if warehouse sizing, auto-suspend, query patterns, and storage are not actively managed. This guide covers the primary cost levers and the monitoring approach that keeps Snowflake spend predictable.
LookML is Looker's proprietary modeling language — the layer that translates raw tables into business-friendly dimensions and measures. This guide covers the core LookML objects, how to structure a LookML project, and the best practices that keep models maintainable.
Every modern data team needs a data quality framework. This guide compares the leading tools — Great Expectations, Soda Core, dbt tests, and Monte Carlo — on coverage, implementation overhead, and the right use case for each.
A single-node Tableau Server is a single point of failure. High availability deployments use multiple nodes across process types to eliminate that risk. This guide covers the HA architecture options, the minimum viable configuration, and the operational requirements.
The complete guide to dimensional modeling — star schemas, snowflake schemas, slowly changing dimensions, and the design decisions that determine whether your data warehouse delivers fast, intuitive analytics.
How to implement data governance in a real organisation — data ownership models, metadata management, data classification, access control, data lineage, and the change management required to make governance stick without killing analytical agility.
The design and development principles that distinguish Tableau dashboards that get used from dashboards that get ignored — layout, colour, typography, performance, user testing, and the governance practices that keep your certified content clean.
What dbt Cloud provides over dbt Core — CI/CD pipelines, the IDE, scheduled jobs, environment management, and the operational features that make the difference between a development tool and a production data platform.
What DuckDB is, why it has become the standard tool for local analytical workloads, how it fits in the modern data stack alongside Snowflake and BigQuery, and the practical use cases where DuckDB outperforms traditional approaches.
A practical comparison of Polars and pandas for data engineering workloads — performance, API differences, memory model, and the specific scenarios where switching from pandas to Polars delivers meaningful improvements versus adding complexity without benefit.
What ClickHouse is, when it beats Snowflake and BigQuery for specific workloads, how its architecture produces sub-second query times on billions of rows, and the use cases where organisations choose ClickHouse over traditional cloud data warehouses.
How Dagster differs from Airflow — asset-based orchestration, software-defined assets, asset materialisation, observability, and the cases where Dagster reduces data pipeline complexity versus cases where Airflow remains the better choice.
A practical comparison of the three major open table formats for data lakes — Apache Iceberg, Delta Lake, and Apache Hudi — covering architecture, use cases, cloud platform support, and how to choose for your specific workload.
What Trino (formerly PrestoSQL) is, how it enables SQL queries across Hive, S3, Snowflake, PostgreSQL, and other sources without data movement, and when federated query makes sense versus ETL into a central warehouse.
How the modern data stack has evolved — what tools have consolidated, what categories have been disrupted, what the AI era changes about data infrastructure, and what the stack looks like for new builds versus established environments in 2025.
A direct comparison of Snowflake and BigQuery across pricing model, performance, ecosystem, governance, and total cost of ownership — with guidance on which platform fits which organisational context.
How to build a comprehensive data quality testing strategy with dbt — built-in generic tests, singular tests, dbt-expectations for advanced assertions, test coverage strategy, and how to structure testing so it catches real data quality failures without creating maintenance overhead.
The BigQuery practices that separate well-run data platforms from expensive, slow ones — partitioning strategy, clustering design, slot management, query cost control, data lifecycle policies, IAM governance, and the patterns that reduce monthly BigQuery bills by 50% or more.
The Redshift practices that determine whether your cluster runs efficiently or expensively — sort key and distribution key design, VACUUM and ANALYZE maintenance, WLM queue configuration, Redshift Serverless vs provisioned, and the query patterns that cause the most performance degradation.
A clear-eyed comparison of data warehouses and data lakes — what each is actually for, where the lakehouse fits, the workloads that belong in each, and the architecture decision framework that prevents organisations from buying the wrong platform for the wrong problem.
The DAX concepts that actually matter in production Power BI models — the difference between row context and filter context, CALCULATE and its many uses, time intelligence patterns, iterating functions, and the performance implications of common DAX patterns.
An honest comparison of Databricks and Snowflake — their architectural differences, the workload types each excels at, how the pricing models compare for analytics and ML workloads, ecosystem and governance trade-offs, and the organisational contexts in which each platform creates more value.
A direct comparison of dbt Cloud and dbt Core across orchestration, development environment, CI/CD, the Semantic Layer, team size fit, and cost — with guidance on the inflection points where Cloud creates enough value to justify the spend.
The Snowflake cost levers that matter — warehouse auto-suspend configuration, warehouse sizing experiments, query result caching, storage compression, search optimisation, and the monitoring queries that surface the highest-cost workloads before the bill arrives.
The design principles that determine whether a dashboard actually gets used — visual hierarchy, chart type selection for specific analytical questions, layout rhythm, performance as a design constraint, and the common patterns that make dashboards look professional but fail to support decisions.
The Airflow patterns that separate production-grade pipelines from fragile ones — idempotent task design, DAG parameterisation, dependency and trigger strategies, scaling configuration, monitoring and alerting, and the common mistakes that cause Airflow environments to degrade over time.
The Tableau development standards that separate workbooks that remain maintainable and performant over time from ones that become impossible to modify six months after they were built — calculation design, data source management, layout standards, documentation, and the governance practices that prevent technical debt accumulation.
The data infrastructure decisions that determine whether a startup builds analytical capability or analytical debt — when to invest in a data warehouse, what to build versus buy at each stage, the premature abstractions to avoid, and the architecture that scales from seed to Series B without a rewrite.
A direct comparison of Prefect and Apache Airflow for data pipeline orchestration — architecture differences, task design philosophy, deployment models, testing approach, and the team contexts where each platform creates more value.
The organisational roles that make data governance work in practice — what a Chief Data Officer actually does, the difference between data owner and data steward, how governance committees function, and the role design that enables accountability without creating bottlenecks.
The core differences between data mesh and data fabric — mesh as a sociotechnical approach decentralising ownership to domain teams versus fabric as a technology layer providing unified access across a centralised architecture — and the organisational and technical factors that determine which pattern fits.
An honest comparison of the three dominant cloud data platforms — Snowflake, BigQuery, and Databricks — across architecture, performance, pricing model, ecosystem, and the organisational and technical factors that should drive the decision for your specific environment.
The five categories of Tableau calculations — basic, aggregate, table, LOD, and parameter-based — when each type is appropriate, the most common mistakes analysts make, and the patterns that separate well-engineered Tableau workbooks from ones that break under scale or confuse the next maintainer.
A plain-language guide to every layer of the modern data stack — ingestion, storage, transformation, orchestration, cataloguing, and BI — the leading tools at each layer, the architectural decisions that determine which tools belong in your stack, and what the modern data stack gets right and wrong.
How to diagnose and fix Tableau Server performance problems — the tools for measuring server health, the configuration parameters that matter, workbook-level performance patterns that cause server load, and the operational practices that prevent performance degradation over time.
Data engineers and analytics engineers both work with data pipelines and SQL, but the roles have distinct scopes, tools, and career paths. This guide draws the line clearly — what data engineers own, what analytics engineers own, where the overlap is, and how the two roles divide responsibilities in organisations of different sizes.
Everything Tableau Cloud site administrators need to manage their environment effectively — site configuration, user and group management, project permissions, content governance, extract and refresh management, and the REST API and admin views that provide operational visibility.
Data observability platforms automatically detect anomalies in data pipelines — unexpected row count drops, freshness failures, schema changes, and distribution shifts — before business users notice. This guide covers the leading tools, what each does well, and how to decide whether to buy a platform or build observability on top of dbt and open-source tooling.
BigQuery charges by bytes scanned, which means every unoptimised query is a cost event. This guide covers the techniques that reduce BigQuery costs — table partitioning, clustering, column selection, BI Engine reservations, slot reservations vs on-demand, query cost governance, and the monitoring setup that makes cost problems visible before they become surprises.
dbt snapshots capture point-in-time historical records for slowly changing dimension data — customer addresses, subscription statuses, account tiers — so you can answer questions about what something was at a specific point in time, not just what it is now. This guide covers when to use snapshots, the snapshot strategy options, and the most common implementation mistakes.
dbt sources define the raw tables that your transformations build on — the ingested data from Fivetran, Airbyte, or custom pipelines. Properly configured sources enable source freshness testing, consistent referencing across models, and clear lineage from raw data through to marts. This guide covers every aspect of dbt source configuration.
How to structure a dbt project that remains navigable and maintainable as it grows — the staging/intermediate/mart layering convention, directory organisation, file naming conventions, how to use subdirectories for domain separation, configuration inheritance through dbt_project.yml, and the project structure decisions that matter most at different stages of growth.
Cloud data warehouse costs scale with usage in ways that are hard to predict without active management. This guide covers the cost levers in Snowflake, BigQuery, and Redshift, the monitoring and alerting setup that makes cost problems visible early, and the governance practices that prevent runaway spend without blocking analytical work.
Snowflake credits accumulate faster than most organisations expect. This guide covers the specific configuration settings, query patterns, and governance practices that control Snowflake costs without restricting analytical work — virtual warehouse sizing, auto-suspend configuration, resource monitors, query profiling, and the monitoring setup that makes cost problems visible before they appear on the bill.
How Fivetran extracts, loads, and normalises data from source systems — the connector architecture, the normalised schema pattern, how incremental syncs work, the log-based CDC approach for database connectors, and the configuration decisions that determine cost, reliability, and schema compatibility.
Tableau actions transform static dashboards into interactive analytical tools — allowing users to click, hover, or select to filter across multiple sheets, navigate to detail views, open URLs, or change parameter values. This guide covers every action type, the most useful design patterns, and the common mistakes that make actions confusing or unreliable.
How to build a CI/CD pipeline for dbt that automatically tests changes before they reach production, enforces code review, and deploys reliably — covering GitHub Actions configuration, the slim CI pattern with state-modified selection, environment management, and the operational practices that make dbt deployments reliable at scale.
The specific roles in a modern data team, what each role actually does, how to write job descriptions that attract the right candidates, the interview questions that separate strong from weak candidates in technical assessments, and the hiring mistakes that leave data teams underpowered or misstructured for years.
Parameters are the most versatile feature in Tableau — they let users dynamically control calculations, filters, reference lines, and chart types. This guide covers every parameter data type, how to wire parameters into calculations and filters, the difference between parameters and filters, and the advanced patterns that make dashboards genuinely interactive.
Running Tableau Server at enterprise scale requires active administration — monitoring backgrounder health, tuning process counts, managing extract refresh queues, diagnosing VizQL performance degradation, and staying ahead of capacity constraints. This guide covers the operational patterns that keep production Tableau Server environments running reliably.
Most BI dashboards fail not because of the data or the tool, but because of design decisions that make the dashboard cognitively expensive to use. This guide covers the principles that separate dashboards that executives actually open from dashboards that collect dust after the first demo.
CDPs have become one of the most overhyped and misunderstood categories in enterprise software. Most organizations buying a CDP already have the data they need — they are buying an integration problem on top of an existing integration problem. This guide clarifies what CDPs actually do, when they solve a real problem, and when they do not.
Data mesh is an architectural pattern that requires organizational change to succeed. The technology decisions — federated data ownership, domain-oriented data products, a self-serve infrastructure platform — are consequential, but they follow from the organizational design. This guide covers the organizational design decisions that determine whether a data mesh implementation succeeds or stalls.
Analytics environments are high-risk targets. They concentrate sensitive data, they are accessed by many users with varying security hygiene, and they often have weaker security controls than the operational systems they source data from. This guide covers the security controls that matter most for protecting analytical data.
Analytics engineering has converged on a set of core tools — dbt for transformation, Git for version control, data warehouses for compute, and orchestration tools for scheduling. But the choices within each category have multiplied. This guide covers the toolchain decisions that matter and how to evaluate the options for your specific context.
The modern data stack — cloud warehouse, dbt, managed ingestion connectors, BI tool — replaced the traditional ETL-to-data-warehouse pattern for good reasons. But it also brought new limitations and failure modes that organisations encounter after the initial implementation succeeds. This is an honest assessment of what the modern data stack is, why it works, and where it falls short.
The choice between a Tableau extract and a live data source connection is one of the most impactful performance decisions in Tableau architecture. It determines data freshness, query response time, infrastructure load, and extract storage costs. This guide covers when each option is appropriate and how to optimise whichever you choose.
Tableau has three deployment options with fundamentally different security, collaboration, and cost profiles. Choosing the wrong deployment creates either over-investment (paying for Server capabilities you do not need) or under-investment (using Public for content that should be secured). This guide clarifies the differences and the right choice for each context.
Tableau Server sizing is one of the most mishandled infrastructure decisions in enterprise analytics. Under-sized servers produce slow dashboards and backgrounder backlogs. Over-sized servers waste infrastructure budget. This guide covers how to size correctly from the start and how to diagnose and fix sizing problems in existing environments.
Redshift and Snowflake are the two most commonly evaluated cloud data warehouses for mid-market enterprise analytics. Both are mature, capable platforms — but they make different architectural trade-offs that matter depending on your workload, team, and cloud strategy. This guide gives you the honest comparison.
Most data pipelines work — until they do not. The difference between pipelines that are reliable, maintainable, and debuggable and pipelines that accumulate technical debt until they fail in production is a set of practices that are easy to skip under deadline pressure and costly to retrofit later.
ETL — extract, transform, load — was the standard data integration pattern for decades. ELT — extract, load, transform — has replaced it in most modern data stacks. The shift is not just a letter swap; it reflects a fundamental change in where and how data transformation happens, with significant implications for architecture, tooling, and team skills.
Tableau Server migrations — version upgrades, environment changes, or Cloud migrations — are higher-risk operations than most organisations anticipate. The workbooks and data sources that work in the current environment do not always behave identically in the target. This guide covers the migration patterns, pre-migration testing, and the failure modes that catch organisations by surprise.
Data lineage answers the question every data team gets asked eventually: where does this number come from? When a metric looks wrong, when a regulatory audit requires proof of data provenance, when a schema change breaks a downstream dashboard — lineage is the map that makes these investigations tractable rather than requiring days of manual tracing.
Tableau Server administration is not glamorous work — but its absence is highly visible when extracts fail overnight, VizQL performance degrades, or a misconfigured upgrade takes down the environment for a business day. This guide covers the operational practices that keep Tableau Server environments stable, performant, and maintainable.
Cloud data warehouse costs have a way of growing faster than the analytical value they deliver. The spend is often justified by capability that exists in theory but is not being used. This guide covers the specific techniques for reducing Snowflake, BigQuery, and Redshift costs without sacrificing the performance and capability you actually depend on.
Tableau makes it easy to put charts on a dashboard. It does not make it easy to design dashboards that communicate clearly, guide the eye to what matters, and feel polished rather than cluttered. This guide covers the design decisions — layout, colour, typography, whitespace — that separate dashboards people trust from dashboards people ignore.
Tableau Server performance issues manifest in ways that are easy to observe and hard to diagnose: dashboards are slow, extracts fail, users complain about load times. The root causes vary widely and require systematic investigation. This guide covers the diagnostic approach that identifies performance bottlenecks at the correct layer.
Tableau's licence tiers — Creator, Explorer, and Viewer — cover a wide range of capability and price. Organisations that assign licences without a clear framework end up either over-licensed (paying for Creator access for users who only need Viewer) or under-licensed (constraining analysts who need Explorer or Creator capability). This guide clarifies the decision framework.
dbt is well-understood at the basic level — staging models, marts, tests, docs. The patterns that separate dbt projects that scale from those that accumulate technical debt are less commonly documented: how to manage incremental models correctly, design generic tests that catch real issues, structure macros without overcomplicating them, and operate dbt in a production environment reliably.
Databricks provides unified analytics on Apache Spark with Delta Lake storage. Getting consistent performance and manageable costs requires deliberate configuration choices at the cluster, storage, and pipeline levels. Default settings are rarely optimal for production workloads.
Most Tableau users operate at a fraction of the tool's capability — not because the advanced features are difficult, but because they are not visible unless you look for them. The features that distinguish experienced Tableau developers from casual users are learnable, practical, and immediately applicable to complex analysis problems.
Most Tableau dashboards are built to show what data exists rather than to answer a specific question. The difference between dashboards that drive decisions and dashboards that get ignored comes down to design principles that can be learned, applied consistently, and evaluated objectively.
Qlik Sense and Tableau are two of the longest-standing enterprise BI platforms. They share the same target market but have fundamentally different analytical philosophies — Qlik is associative and data-model-centric, Tableau is visual-first and exploration-centric. The right choice depends on the type of analysis your organisation needs most.
Tableau Prep is the data preparation tool in the Tableau platform — purpose-built for cleaning, reshaping, and combining data before it reaches Tableau Desktop or Tableau Cloud for visualisation. It gives analysts a visual, step-by-step interface for data transformation that produces reproducible, shareable preparation flows without requiring SQL expertise or engineering support.
The data lakehouse combines the storage economics and flexibility of a data lake with the query performance and governance of a data warehouse. For many organisations, the question is no longer which architecture to use but which use cases are better served by managed warehouse services (Snowflake, BigQuery, Redshift) and which by open lakehouse formats (Iceberg, Delta Lake, Hudi) on object storage.
BigQuery's on-demand billing model charges per byte scanned. Without careful query and table design, costs can grow unexpectedly — a single unpartitioned query on a multi-terabyte table generates significant spend. This guide covers the BigQuery cost optimisation patterns that teams use to control spend: partitioning, clustering, materialised views, query governance, and the evaluation framework for reserved slots versus on-demand billing.
Migrating from a legacy on-premises data warehouse — Teradata, SQL Server, Oracle — to Snowflake is a multi-phase programme covering data migration, SQL dialect conversion, ETL pipeline rewiring, BI tool reconnection, and user transition. This guide covers the full migration lifecycle, common blockers, and how to structure the programme to minimise business disruption.
Looker uses LookML — a modelling language that sits between your data warehouse and end-user queries — to define dimensions, measures, and joins in version-controlled YAML-like files. This guide covers LookML Explores, views, joins, dimension and measure types, derived tables, and how the modelled semantic layer enables governed self-service analytics.
SQL joins combine rows from two or more tables based on a related column. This guide explains every join type — INNER, LEFT, RIGHT, FULL OUTER, CROSS, and SELF — with diagrams, concrete examples, and the analytical use cases where each join type is appropriate. Covers common join mistakes and performance implications.
Common Table Expressions (CTEs) let you define named subqueries at the top of a SQL statement and reference them like temporary tables throughout the query. This guide covers CTE syntax, recursive CTEs, when to use CTEs versus subqueries or temporary tables, and patterns for using CTEs to build readable multi-step analytical queries.
Most dashboards fail not because of bad data, but because of design decisions that obscure rather than reveal insight. This guide covers the core principles of effective dashboard design — audience and purpose definition, information hierarchy, chart selection, layout and visual flow, and the common mistakes that produce dashboards that look impressive but are never used.
Chart selection is a decision about what comparison to make visible. A bar chart answers a different question than a line chart; a scatter plot reveals patterns invisible in a table. This guide covers which chart types to use for which analytical purposes, common visualisation mistakes that distort data, and design principles for charts that communicate accurately.
SQL aggregation collapses multiple rows into a single summary value — COUNT, SUM, AVG, MIN, MAX. This guide covers GROUP BY fundamentals, multi-column grouping, HAVING for filtering aggregated results, GROUPING SETS and ROLLUP for multi-level summaries, and common aggregation mistakes that produce incorrect results.
Analytics engineering and data analysis are related but distinct disciplines. Data analysts turn data into insight; analytics engineers build the data infrastructure that makes analysis possible. This guide clarifies the distinction, describes how the two roles interact, explains the skills each requires, and helps organisations understand when they need which — or whether one person can do both.
OLTP (Online Transaction Processing) and OLAP (Online Analytical Processing) are the two fundamental database patterns — one optimised for recording individual transactions, the other for querying and analysing large volumes of historical data. This guide explains the difference, why it matters for system design, and how modern data architectures handle both.
Tableau and Power BI are the two dominant enterprise BI platforms, and choosing between them is a significant investment decision. This guide compares them honestly — strengths, weaknesses, cost, ecosystem, governance, and the organisational contexts where each genuinely excels — without the vendor-driven framing that dominates most comparisons.
Snowflake is a cloud-native data platform built on a multi-cluster shared data architecture — separating compute from storage, enabling elastic scaling, and supporting multiple independent compute clusters querying the same data simultaneously. This guide explains how Snowflake works, what makes it different from traditional data warehouses, and the use cases where its architecture is a strong fit.
Databricks is a cloud-native data and AI platform built on Apache Spark — providing managed Spark clusters, collaborative notebooks, Delta Lake open table format, Unity Catalog for data governance, and a SQL warehouse for BI queries. This guide explains what Databricks is, how its lakehouse architecture works, what the platform includes, and when it is the right choice versus a managed data warehouse.
A database and a data warehouse both store data in tables and respond to SQL queries — but they are designed for fundamentally different purposes. This guide explains the difference between an operational database and an analytical data warehouse, when you need each, and why querying your production database for analytical reports is a common mistake with predictable consequences.
Data mesh is a decentralized approach to data architecture that treats data as a product owned by domain teams rather than managed centrally by a data engineering team. This guide explains the four principles of data mesh, when it makes sense, and the organizational challenges that stop most implementations.
Medallion architecture organizes a data lakehouse into progressive quality layers: bronze (raw ingested data), silver (cleaned and validated), and gold (business-ready aggregations). This guide explains how each layer works, what belongs in each, and when the pattern adds value over simpler alternatives.
ETL is a specific pattern for moving data. A data pipeline is a broader term for any automated workflow that moves or transforms data. Understanding the distinction — and when each term applies — helps clarify how modern data stacks are actually organized.
The schema design of a data warehouse determines how fast analytical queries run, how easy models are to maintain, and how well BI tools can auto-generate SQL. This guide explains the three primary patterns — star schema, snowflake schema, and wide denormalized tables — with the trade-offs that determine when each is appropriate.
A data warehouse migration moves an organization's analytical data environment from one platform to another — from on-premise to cloud, or between cloud warehouses. This guide explains the key phases of a warehouse migration, the common failure points, and the architectural decisions that determine whether the migration delivers its intended value.
Dashboards and reports are both analytical outputs, but they serve different purposes, audiences, and decision-making contexts. Understanding the distinction helps teams build the right artifact for each use case and avoid building dashboards when reports are needed or vice versa.
Data warehouse architecture defines how data flows from source systems through transformation layers to analytical consumption. This guide explains the canonical layers of a modern warehouse, common design patterns, and the decisions that shape warehouse architecture in practice.
dbt Cloud is the managed deployment platform for dbt (data build tool), adding orchestration, a web IDE, documentation hosting, job scheduling, and collaboration features to the open-source dbt Core framework. This guide explains what dbt Cloud provides, how it compares to self-hosted dbt, and when each is appropriate.
A data fabric is an architectural approach that provides unified data access, integration, and governance across distributed, heterogeneous data sources — without physically centralizing them. This guide explains what data fabric means in practice, how it differs from a data mesh, and when the concept provides genuine value.
Tableau Server administration is the discipline of maintaining, monitoring, and optimizing a Tableau Server deployment — from initial installation through ongoing performance management, user governance, and upgrade planning. This guide explains the core responsibilities and what separates reactive from proactive administration.
Tableau Prep is Tableau's data preparation tool, allowing analysts to clean, reshape, and combine data before loading it into Tableau Desktop or publishing to Tableau Server. This guide explains what Tableau Prep does, where it fits in the analytics workflow, and when it is and is not the right tool.
A data governance council is the cross-functional body that makes decisions about data standards, policies, ownership, and priorities in an organization. This guide explains how governance councils are structured, what they decide, and why they succeed or fail.
Snowflake and BigQuery are the two most widely deployed cloud data warehouses. This guide compares them on architecture, pricing, performance, ecosystem fit, and governance — and explains how to choose based on your organization's specific requirements.
Cloud data warehouse costs can grow rapidly as data volumes, query workloads, and team usage expand. This guide explains the primary cost drivers in Snowflake, BigQuery, and Redshift, and the optimization strategies that reduce spend without degrading analytical capability.
The modern data stack is the collection of cloud-native tools that replaced legacy on-premises data infrastructure — covering ingestion, storage, transformation, orchestration, and BI. This guide explains what the modern data stack is, how the tools fit together, and what distinguishes it from the previous generation.
Tableau Cloud (formerly Tableau Online) is the fully managed, SaaS version of Tableau's analytics platform. This guide explains what Tableau Cloud provides, how it differs from Tableau Server, what the migration from Server to Cloud involves, and the governance considerations for cloud-hosted analytics.
Dashboard design in Tableau determines whether analytical data communicates clearly or creates confusion. This guide explains the design principles — layout, visual hierarchy, chart selection, color, and interactivity — that distinguish dashboards that drive decisions from dashboards that are technically correct but analytically unhelpful.
Tableau extracts are local copies of your data stored in the columnar Hyper format — the primary mechanism for making dashboards fast and independent of source system availability. This guide explains how extracts work, when to use them over live connections, and how to manage extract refresh schedules.
Master data management (MDM) is the practice of creating and maintaining a single, authoritative record for core business entities — customers, products, locations, and accounts — across all systems in an organization. This guide explains why MDM matters, how it is implemented, and what it costs to get wrong.
Power BI is Microsoft's cloud-connected business intelligence platform for creating dashboards, reports, and data models from enterprise data sources. This guide explains how Power BI works, where it fits in the Microsoft data ecosystem, and how it compares to Tableau for enterprise analytics.
Tableau Server is Tableau's self-hosted enterprise analytics platform — the on-premises or private cloud deployment of Tableau that organizations install and manage themselves. This guide explains Tableau Server's architecture, capabilities, administrative responsibilities, and how it compares to Tableau Cloud.
Tableau Desktop is the Windows and Mac authoring application where analysts and data professionals build Tableau workbooks — connecting to data sources, creating visualizations, and designing dashboards. This guide explains Tableau Desktop's capabilities, workflow, and how it fits in the broader Tableau platform.
Tableau parameters are workbook variables that allow dashboard users to input values that dynamically change calculations, filters, and reference lines. This guide explains how Tableau parameters work, common use cases, and how they differ from filters.
Tableau actions are interactive behaviors that connect views on a dashboard — filtering, highlighting, navigating, and triggering URLs based on user selections. This guide explains the four types of Tableau actions, how they work, and how to use them to build intuitive dashboard navigation and drill-down experiences.
A Tableau set is a custom subset of dimension members defined by a condition, manual selection, or top N filter. Sets enable in/out comparisons and can be used across multiple views to create consistent comparative analysis. This guide explains how sets work, how they differ from groups and filters, and common analytical use cases.
Modern data warehouses are organized in layers — raw ingested data, cleaned and standardized data, and business-ready analytical data — each serving a specific purpose in the transformation pipeline. This guide explains the layered architecture, the medallion architecture pattern, and how dbt implements it.
Tableau row-level security (RLS) restricts which rows of data each user can see when they view a dashboard, without creating separate workbooks for each audience. This guide explains the three main implementation patterns — user filters, entitlement tables, and data source permissions — and how to choose between them.
Data warehouse costs depend on storage, compute, and data transfer usage across cloud platforms. This guide explains how Snowflake, BigQuery, and Redshift are priced, the common cost drivers, and the strategies data teams use to reduce costs without sacrificing analytical capability.
A data fabric is an architecture that provides consistent access to data wherever it lives — across cloud platforms, on-premises systems, and data lakes — without requiring centralized physical consolidation. This guide explains how data fabrics work, what problems they solve, and when they are (and are not) the right architectural choice.
Tableau LOD (Level of Detail) expressions let you compute aggregations at a different granularity than the view — answering questions like "what is the first purchase date per customer regardless of how the view is filtered." This guide explains FIXED, INCLUDE, and EXCLUDE LOD expressions with practical use cases.
Tableau tooltips display contextual information when users hover over marks in a visualization. Custom tooltips can include calculated fields, additional dimensions, and even embedded visualizations — making them one of the most effective tools for providing detail-on-demand without cluttering the main view.
dbt sources are declarations of the raw tables and views in your data warehouse that your dbt project reads from. Declaring sources in dbt enables source freshness testing, lineage documentation, and schema assertions against the raw layer — the foundation of a robust dbt project.
Tableau maps visualize data geographically — by country, state, city, zip code, or custom geographic boundaries. This guide explains the different map types in Tableau, when to use geographic visualization, how to work with custom shapes and spatial files, and performance considerations for map-heavy dashboards.
dbt macros are reusable blocks of SQL logic written with Jinja templating that can be called across multiple models in a dbt project. Macros eliminate repetition in SQL transformation code, enable dynamic query generation, and let dbt projects share logic via packages.
dbt seeds are CSV files committed to your dbt project repository that dbt loads directly into your data warehouse as tables. They provide a managed, version-controlled way to maintain small, static reference datasets — country codes, product categories, cost center mappings — without requiring an EL pipeline.
Tableau data blending is a method of combining data from two different data sources in a single view without a database-level join. Instead of merging data in the database, Tableau queries each source independently and combines results at the visualization layer. This guide explains how blending works, when it is appropriate, and its key limitations.
dbt exposures are declarations in your dbt project of the downstream consumers of your dbt models — Tableau dashboards, Power BI reports, ML pipelines, and applications that read from tables your dbt project produces. Declaring exposures makes the full end-to-end lineage visible, from source data through transformation to the dashboards business users see.
A Tableau Prep Flow is a visual data preparation workflow built in Tableau Prep Builder that cleans, shapes, and combines data before it reaches Tableau Desktop or Tableau Server for visualization. This guide explains how Prep flows work, when they are the right tool, and how they fit into a governed analytics architecture.