1713
Bernoulli publishes the law of large numbers
Ars Conjectandi introduced a result connecting repeated observations with long-run frequency. It established a mathematical bridge between chance models and empirical regularity.
← Data Analyst career guide
Data and AI
Data Analyst timeline
How the field of a Data Analyst got here: the sourced milestones from every published course of this career, on one axis.
1713 to 2025
182 milestones across 20 courses
- Core
- Supporting
- Specialization
The axis stretches toward recent years, where most milestones fall. Bars count milestones per year; dots are single milestones.
Before 1990
56 milestones
1763
Bayes's essay introduces inverse probability
The posthumous essay gave a rule for updating a probability after observing evidence. It became the basis for the prior, likelihood, and posterior structure used in Bayesian inference.
1786
Playfair publishes The Commercial and Political Atlas
William Playfair used statistical graphics to present commercial and political data. The work established charts as a form for comparing quantitative change and magnitude.
1805
Legendre publishes the method of least squares
Legendre's appendix on determining comet orbits introduced the least squares criterion for combining observations, the estimation rule every regression fit still uses.
1809
Gauss connects least squares to probability theory
Gauss's Theoria motus developed the method independently and grounded it in the theory of errors, giving least squares the probabilistic foundation that underpins regression inference.
1812
Laplace publishes Théorie analytique des probabilités
Laplace consolidated probability methods and their applications to observational error and inference. The work helped establish probability as a general analytical tool rather than a gambling topic.
1855
Snow's cholera map appears in the St James report
The report included a map of cholera cases around the Broad Street pump. It made spatial distribution part of an argument about a public-health question.
1869
Minard publishes the map of Napoleon's Russian campaign
Charles Joseph Minard combined geographic position, army size, direction, and temperature in one statistical graphic. It remains a reference for combining several variables without treating a chart as decoration.
1886
Galton names "regression" from heredity studies
Galton's analysis of hereditary stature observed offspring heights regressing toward the population mean, and the term became the name of the modeling method that grew out of that work.
1900
Pearson introduces the chi-squared goodness-of-fit test
Pearson's criterion compared observed and expected frequencies. It supplied a reusable test statistic for categorical data and model fit.
1908
Gosset publishes Student's t distribution
Working as "Student" at Guinness, Gosset derived the exact small-sample behavior of estimates, giving regression coefficients their correct standard errors and intervals at realistic sample sizes.
1922
Fisher formalizes calculation of P from contingency tables
Fisher's paper treated the interpretation of chi-squared statistics and calculation of P. It shaped the significance-testing vocabulary later used in applied hypothesis testing.
1922
Fisher formalizes the theory of statistical estimation
Fisher's paper on the mathematical foundations of theoretical statistics established the estimation framework, including consistency and efficiency, that regression's optimality claims rest on.
1924 May
Shewhart proposes the control chart
Walter Shewhart's Bell Laboratories memorandum introduced a chart for separating stable process variation from signals of special causes.
1926
Fisher publishes The Arrangement of Field Experiments
The paper set out rational principles for field experiments, including randomization and valid estimation of error. These principles became foundations for later controlled comparisons.
1932
Likert publishes a technique for measuring attitudes
Rensis Likert showed that a summed set of agree-disagree items could measure a social attitude reliably, using a simple scoring method instead of the elaborate judging procedure it replaced. The multi-item ordinal scale became the standard way to quantify opinion strength in questionnaires.
1933
Neyman and Pearson publish the testing framework
Their work defined a decision framework for testing statistical hypotheses. It made error rates and power central design quantities.
1934
Neyman formalizes probability sampling
Jerzy Neyman's paper to the Royal Statistical Society compared stratified random sampling with purposive selection and gave random sampling a formal basis, including confidence intervals for the resulting estimates. It established why known selection probabilities are required for valid inference.
1935
Fisher publishes The Design of Experiments
Fisher's book presented experimental procedure and significance testing together. It became a reference point for randomization, replication, and planned inference.
1936 Nov
The Literary Digest poll fails and scientific polling gains credibility
The Literary Digest mailed 10 million straw ballots drawn from car and telephone owners and predicted the wrong presidential winner, while Gallup, Roper, and Crossley called it correctly from far smaller quota samples. Later analysis attributed the failure to a biased frame combined with low, self-selected response.
1946 Jun
Plackett and Burman publish multifactorial designs
The paper described optimum multifactorial designs for studying many component changes against a measured characteristic. It provided a compact screening approach when a full factorial is too large.
1947 Sep
AAPOR is founded
The American Association for Public Opinion Research was formally organized on September 4, 1947, following a 1946 conference in Central City, Colorado. It became the body that would later publish shared standards for disclosure, response-rate reporting, and survey practice in the United States.
1948
The Truman-Dewey polling error prompts a formal review
All three national pollsters predicted the wrong winner in the 1948 US presidential election, having stopped interviewing early and relied on interviewer-controlled quota sampling. The Social Science Research Council's committee report faulted quota selection and premature fieldwork, accelerating the shift to probability samples.
1948 Oct 30
The MRC streptomycin trial is published
The British Medical Journal published a controlled tuberculosis trial with concealed random allocation. The design demonstrated randomized comparison in a high-stakes medical setting.
1950 Mar
Grubbs publishes outlying-observation criteria
Frank Grubbs formalized statistical criteria for testing suspected outlying observations in a sample.
1951 Jan
Box and Wilson publish response-surface methodology
Box and Wilson described sequential experiments for finding settings that optimize a quantitative response. The method made experimentation an iterative route toward operating conditions rather than a single comparison.
1957
Holt and Winters formalize exponential smoothing
Holt described exponential smoothing with a trend term, and Winters extended it to seasonality in 1960. The method forecasts a weighted average of past values with weights that decay exponentially, and it remains a default for clean univariate series.
1958
Luhn publishes A Business Intelligence System
Hans Peter Luhn published A Business Intelligence System. It explored using technology to gather intelligence and provided an early technical framing for information that guides action.
1967
Bertin publishes Semiology of Graphics
Jacques Bertin described visual variables and their use in graphic representation. The work gave later visualization practice a vocabulary for marks and visual encodings.
1970
Box and Jenkins publish the ARIMA methodology
Time Series Analysis: Forecasting and Control set out the identify-estimate-diagnose cycle for autoregressive integrated moving-average models, making autocorrelation-based modeling a systematic procedure rather than an art.
1970
Codd defines the relational model
E. F. Codd published A Relational Model of Data for Large Shared Data Banks. The model established the relational structures and operations that later became the basis for analytical data storage and SQL querying.
1970
Codd publishes the relational model
E. F. Codd published A Relational Model of Data for Large Shared Data Banks. It established the table-and-relation model that SQL later made usable as a query language.
1970
Tukey circulates a preliminary Exploratory Data Analysis edition
Tukey's preliminary edition established exploratory analysis as a distinct approach before its later book publication.
1973
Anscombe shows identical fits, different data
Anscombe's constructed quartets produced the same fitted lines, R-squared, and intervals from visibly different relationships, making residual plotting a mandatory step rather than an optional extra.
1976
Rubin formalizes inference with missing data
Rubin distinguished mechanisms that generate missing values and their consequences for inference. Missingness became a modeling and study- design issue rather than a data-cleaning afterthought.
1977
Exploratory Data Analysis is published
Tukey's book collected graphical and resistant techniques under the EDA name, giving practitioners a shared vocabulary for probing data before confirmation.
1977
Tukey publishes Exploratory Data Analysis
John Tukey presented exploratory data analysis as a way to use displays to investigate data before formal conclusions. This reinforced the distinction between exploratory and communicative views.
1978
Dillman publishes the Total Design Method
Don Dillman's book set out a detailed, standardized procedure for mail and telephone surveys, covering contact sequence, wording, layout, and follow-up. Applied as a whole, it raised typical mail response rates to the 60 to 70 percent range and made survey administration a documented method rather than local craft.
1979
Efron introduces bootstrap methods
The bootstrap used resampling to estimate uncertainty when a closed- form sampling distribution is inconvenient. It made computational inference practical across a wider range of estimators.
1979
Oracle V2 commercializes SQL
Relational Software, Inc. introduced Oracle V2 as the first commercial SQL implementation. Commercial relational databases made a standardized storage and query foundation available to enterprise analytical systems.
1979
VisiCalc ships on the Apple II
Bricklin and Frankston's VisiCalc made recalculating tables interactive on a personal computer, turning typed assumptions into live formula results and creating the spreadsheet modeling habit.
1979 May
VisiCalc reaches personal computers
VisiCalc was publicly demonstrated in 1979 and established the electronic spreadsheet as a practical personal-computer application.
1980 Apr
Taguchi publishes the loss function for quality variation
Taguchi related deviation from a target to loss rather than treating every in-specification result as equally good. This sharpened the practical focus on reducing variation in designed processes.
1982
The first M-competition standardizes forecast evaluation
Makridakis and collaborators forecast a large set of real series with many methods and scored them out of sample. The result, that statistically simple methods matched or beat complex ones, set the empirical, benchmark-driven culture the field still uses.
1983
Lotus 1-2-3 dominates the IBM PC
Lotus 1-2-3 combined spreadsheet calculation with graphing and database-like features on the IBM PC, making larger planning models practical for business users beyond the Apple II niche.
1983
Tufte publishes The Visual Display of Quantitative Information
Edward Tufte collected principles and examples for displaying quantitative information. The book made graphical integrity and information-dense displays central topics in visualization design.
1983 Jan
Lotus 1-2-3 combines calculation, data, and charts
Lotus 1-2-3 brought spreadsheet, database, and graphing functions to the IBM PC, making the grid a business-analysis workspace.
1985
Excel 1 includes VLOOKUP on Macintosh
Microsoft later documented that VLOOKUP was included in Excel 1 for Macintosh, anchoring the lookup pattern taught in this course.
1985
Excel launches on Macintosh
Microsoft Excel brought a graphical spreadsheet with richer formatting and formula tooling on the Mac, establishing the product line that later carried mainstream what-if features for Windows users.
1985
Stata 1.0 is released
Stata shipped its first version in January 1985, joining a statistical-software market that put least squares fits and their diagnostics on professional desktops rather than mainframes.
1986
ANSI publishes SQL-86
ANSI published the first SQL standard. A standard core gave database products a common reference point even as their dialects continued to differ.
1987
Excel for Windows ships
Excel on Windows positioned Microsoft's spreadsheet as the corporate planning surface where scenario-style assumption changes became a daily finance and operations practice.
1987
ISO publishes ISO/IEC 9075
ISO adopted SQL as ISO/IEC 9075. The paired ANSI and ISO work made SQL an international database-language standard.
1987 Oct
Excel for Windows ships
Microsoft announced Excel for Windows in October 1987 and shipped it in November, extending Excel's spreadsheet workflow to Windows.
1989
Business intelligence becomes a decision-support umbrella term
Howard Dresner used business intelligence for concepts and methods that improve business decisions with fact-based support systems. The term brought data warehousing, reporting, and analysis into a recognizable practice area.
1989
Groves codifies total survey error
Robert Groves's Survey Errors and Survey Costs organized coverage, sampling, nonresponse, and measurement error into one mean-squared-error structure and weighed each against the cost of reducing it. The framework gave designers a common vocabulary for trading error sources against a fixed budget.
1990s
11 milestones
1991
Python becomes publicly available
Python was created in the early 1990s as a successor to ABC. Its early releases established the language that later became community-maintained.
1991 Feb
Python is posted to Usenet
Python's first public release established the language that later became the host environment for numerical and tabular libraries.
1992
SQL-92 expands the standard
ANSI and ISO published the revision commonly called SQL-92. It became a long-lived baseline for the portable syntax developers still recognize.
1992 Jan
Balanced Scorecard introduced in Harvard Business Review
Kaplan and Norton presented a framework that balanced financial measures with customer, internal-process, and learning perspectives. Executive reporting gained a broader structure than financial results alone.
1993
Building the Data Warehouse is published
Bill Inmon published Building the Data Warehouse. It helped establish data warehousing as a separate analytical environment rather than a copy of an operational system.
1993
Early R binaries are published on StatLib
Early public R distributions made an open statistical-computing environment available for interactive analysis and graphics.
1995
R source is released as free software
Releasing R under the GPL created an open route for extending statistical and graphical analysis workflows.
1996
The Data Warehouse Toolkit advances dimensional modeling
Ralph Kimball published The Data Warehouse Toolkit. Its dimensional-design techniques gave practitioners a practical structure for facts, dimensions, grain, and query-centered decision support.
1998
AAPOR issues the first Standard Definitions
AAPOR published a common set of case disposition codes and response-rate formulas so that rates could be compared across surveys and modes. Reporting which definition was used, rather than a bare percentage, became an expectation of transparent survey practice.
1999
One-class support vector novelty detection appears
The method estimated a boundary around an unlabeled reference distribution so new observations outside it could be treated as novel.
1999
SQL:1999 adds object and procedural features
SQL:1999 added features including recursive queries and standardized procedural extensions. It widened SQL beyond basic table queries while retaining the relational core.
2000s
30 milestones
2000
R 1.0.0 is released
R reached version 1.0 in February 2000, consolidating an open statistical environment whose lm fits and formula interface became a default route to regression in research and teaching.
2000 May
Local Outlier Factor is published
LOF measured how an observation's local density compares with its neighbors, allowing different regions to have different normal densities.
2000 Jun
PEP 1 defines the proposal process
PEP 1 established the Python Enhancement Proposal process. Language and community proposals gained a documented route for discussion and decision.
2000 Oct
Python 2.0 is released
Python 2.0 added list comprehensions and cyclic garbage collection. Both changes became part of the practical language baseline for later Python users.
2001
IPython begins as an interactive Python environment
IPython established an interactive Python workflow that later supplied the notebook lineage used for iterative data exploration.
2002
A state space framework unifies exponential smoothing
Hyndman and colleagues organized exponential smoothing into the ETS taxonomy of Error, Trend, and Seasonal components, which gave each method a model, a likelihood, and prediction intervals rather than a point recipe.
2003
Matplotlib enters its early public era
Matplotlib's original 2003-2008 project identity marks the emergence of a Python plotting layer for array-based analysis.
2003
SQL:2003 standardizes window functions
SQL:2003 added window functions and SQL/XML work. Window functions made ranking and calculations across related result rows part of standard SQL practice.
2003
Tableau founded from the VizQL research project
Tableau commercialized a visual query approach developed at Stanford. Interactive visual analysis became more accessible to business users.
2003
Tableau is founded from Stanford research
Tableau was founded from a computer science project that developed VizQL to translate drag-and-drop visual expressions into data queries.
2003
Tableau turns visual querying into a commercial BI product
Tableau was founded to turn the Polaris research prototype into Tableau Desktop. Its VizQL approach translated drag-and-drop visual interactions into data queries, widening access to interactive analysis.
2005
NumPy is created
NumPy combined earlier Numeric and numarray work into the array library that provides the vectorized numerical base used by pandas.
2005
Wilkinson publishes The Grammar of Graphics
Leland Wilkinson described graphics as composable elements rather than a fixed menu of chart types. This model informed declarative charting systems that separate data, transformations, marks, scales, and guides.
2006
Information Dashboard Design published
Stephen Few's book formalized dashboard design around rapid monitoring and effective visual communication. It gave practitioners a focused reference for reducing clutter and strengthening comparisons.
2006 Mar
Microsoft forms the Experimentation Platform team
Microsoft formed ExP as an incubation project for controlled experiments across online properties. Experimentation began moving from isolated analyses toward a shared service.
2006 Jun
Google Spreadsheets limited test announced
Google's hosted spreadsheet made multi-user editing of shared models routine, shifting some scenario discussion into browser collaboration even when desktop what-if menus differed.
2006 Jun
Google starts a limited Google Spreadsheets test
Google announced a limited Google Spreadsheets test, bringing a browser-hosted spreadsheet into the emerging Google Docs service.
2007
Excel adopts the XML-based XLSX default format
Excel 2007 used XLSX as its default XML-based workbook format, making the workbook package a documented interchange format.
2007
Excel defaults to Office Open XML workbooks
The xlsx-centered Excel 2007 generation changed how models are stored and exchanged, while the Ribbon reorganized analysis entry points analysts use to reach what-if workflows.
2007
Google publishes a controlled-experiment case study for advertisers
Google's case study used a statistically designed experiment to assess new advertising features. It documented controlled experimentation as a product-development evaluation method at online scale.
2007
scikit-learn begins as a Google Summer of Code project
The project that became scikit-learn started in 2007 and grew into the Python workbench whose linear-model module sits beside regularized variants of least squares.
2008
pandas development begins
Development started at AQR Capital Management, introducing the project that made labeled DataFrames central to Python data work.
2008
The forecast package brings automatic order selection to R
Hyndman and Khandakar released auto.arima and ets in the R forecast package, so choosing ARIMA orders or an ETS form became a function call driven by information criteria and unit-root tests.
2008 Jul 30
Controlled experiments on the web survey is published online
The survey organized web experimentation terminology, architecture, statistical analysis, and practical pitfalls into a common reference. Its 2009 journal issue connected industrial practice with experimental design.
2008 Dec
Isolation Forest reframes anomaly detection
Isolation Forest isolated anomalies with random feature splits rather than first constructing a profile of normal observations.
2008 Dec
Python 3.0 is released
Python 3.0 made incompatible language and library changes rather than preserving every Python 2 behavior. Porting became an explicit concern for projects spanning the transition.
2008 Dec 11
WCAG 2.0 becomes a W3C Recommendation
WCAG 2.0 established durable web accessibility requirements that also apply to browser-delivered reports, controls, charts, and status cues.
2009
pandas becomes open source
Opening pandas development made its DataFrame and I/O workflow available to the wider Python analysis community.
2009 Nov
PowerPivot previews become available
Microsoft released PowerPivot previews for Excel 2010 and SharePoint 2010, combining in-memory personal modeling with managed browser sharing.
2009 Dec 1
Asynchronous tracking launches
Google introduced an optional asynchronous tracking snippet to reduce tracking-code effects on page loading and avoid script-order dependencies. Collection implementation became less coupled to synchronous page execution.
2010s
59 milestones
2010
scikit-learn's first public release
The February 2010 release put linear regression, ridge, and lasso behind one fit-and-predict interface, embedding least squares as the baseline estimator of the Python machine-learning stack.
2010
statsmodels released after Google Summer of Code 2009
The scipy.stats models code was corrected, tested, and released as statsmodels, giving Python an inference-focused OLS implementation with the coefficient tables and diagnostics regression teaching relies on.
2010 Feb
Tableau Public launches
Tableau 5.1 added a free service for publishing interactive visualizations and public data to the web, creating a public learning and storytelling surface.
2010 Jul
Google publishes overlapping experiment infrastructure
Google described infrastructure that allowed many concurrent experiments while separating parameter namespaces. The paper treated tooling and education as parts of experimentation at scale.
2010 Sep
AAPOR's Cell Phone Task Force addresses the mobile shift
As households dropped landlines, the task force set out how random-digit-dial telephone surveys had to add cell phone frames, handle dual-frame overlap, and manage the higher cost and different coverage of mobile interviewing. Telephone survey design was reworked around a frame that no longer matched the population.
2011
SQL:2011 adds temporal features
SQL:2011 added temporal database features. The standard gained a common way to express time-aware data and queries.
2011 Feb
D3 version 1.0 is released
D3 made data-driven document manipulation available as a web visualization library. It enabled authors to bind data to web standards while retaining direct control over marks and interaction.
2011 Apr 20
Google Analytics version five becomes broadly available
The new interface brought configurable dashboards, event goals, row comparison, and reorganized reporting to all supported languages. Analysis work moved into a redesigned reporting surface.
2011 Sep 29
Realtime reports launch
Realtime reports exposed recent site activity as it happened. Practitioners gained a faster way to smoke-test campaign tagging and observe immediate traffic.
2012 Jun
pandas 0.8 improves time-series handling
pandas 0.8 adopted NumPy datetime64 for time-series representation, improving joins, memory use, and serialization around dated data.
2012 Oct 29
Universal Analytics is announced
Universal Analytics introduced Measurement Protocol and broader custom-data integration for cross-platform measurement. Collection expanded beyond browser page traffic toward devices and organizational data.
2013 Feb
CUPED variance reduction is introduced
Researchers described using pre-experiment data to reduce metric variance in online controlled experiments. The method improved sensitivity without replacing randomized assignment.
2013 Apr
Herndon, Ash, and Pollin document a Reinhart-Rogoff Excel error
Their replication showed that an Excel coding error omitting countries changed published average growth figures, making range integrity a public example of spreadsheet model risk beyond missing what-if menus.
2013 Jul
Power BI for Office 365 is introduced
Microsoft named a cloud self-service BI offering that combined Excel modeling and visualization tools with Office 365 publishing and collaboration.
2013 Jul 8
Microsoft details Power BI for Office 365
Microsoft described browser and mobile access to reports built from Excel-based modeling and visualization components. Cloud distribution became part of the mainstream BI design path.
2014
Pew builds a probability-based online panel
Pew Research Center recruited the American Trends Panel through a national random-digit-dial survey, then moved members to self-administered web interviewing. Probability recruitment with online data collection became a practical alternative to repeated telephone samples.
2014
Project Jupyter begins
Jupyter grew from IPython to support interactive data science and scientific computing across programming languages.
2014
Project Jupyter is born from IPython
Jupyter broadened the notebook model beyond Python for interactive scientific computing and data exploration.
2014 May 30
Enhanced Ecommerce enters beta
Enhanced Ecommerce added structured measurement for product views, cart actions, checkout steps, purchases, refunds, and merchandising. Ecommerce funnels became a first-class analytics implementation.
2015
EuSpRIG restates the spreadsheet-error evidence base
Panko's EuSpRIG paper summarized that cell error rates stay material at workbook scale, reinforcing inspection discipline for operational planning models that rely on scenarios and sensitivities.
2015
Vega presents a declarative visualization grammar
Vega defined a JSON-based grammar for describing visual appearance and interactive behavior. It made data, transformations, scales, guides, and marks explicit parts of a specification.
2015 Apr
Tableau 9 adds Data Interpreter and analytical features
Tableau 9 introduced integrated cleanup for cross-tab spreadsheets along with changes to maps, analytics, and query performance.
2015 Jul
The new Power BI reaches general availability
The Power BI service and Power BI Desktop established the current split between local report authoring and cloud distribution.
2015 Jul 24
The new Power BI service and Desktop reach general availability
Microsoft released the new Power BI service and Power BI Desktop. The release combined cloud-hosted distribution with a desktop authoring tool for reports and dashboards.
2015 Jul 24
The new Power BI service reaches general availability
Microsoft released the redesigned service and Power BI Desktop for general use. Reports, dashboards, semantic models, and custom visuals formed a broadly available reporting stack.
2015 Sep
Python 3.5 adds async and await syntax
Python 3.5 added native coroutines and the async and await syntax. Asynchronous I/O became a first-class language-level pattern.
2016
ASA issues its statement on statistical significance and p-values
The American Statistical Association clarified that a p-value measures compatibility of data with a specified model, not the probability that a hypothesis is true or the size or importance of an effect.
2016
SQL:2016 adds JSON support
SQL:2016 incorporated JSON-related features. Relational databases could standardize some work with semi-structured values without abandoning tables, constraints, and joins.
2016 Feb
Enterprise gateway reaches general availability
The managed gateway gave the Power BI service a centralized route to supported on-premises sources for refresh and live connections.
2016 May
Pew publishes an accuracy assessment of online opt-in samples
Pew compared nine online nonprobability samples against government benchmarks and found accuracy varied widely by vendor and was worst for estimates about racial minorities and other subgroups. The study set expectations for where modeled error in opt-in data is largest.
2016 Aug
Tableau 10 expands cross-database analysis
Tableau 10 added cross-database joins, data-source filtering, clustering, and revised authoring defaults for analysts combining more varied sources.
2017
Vega-Lite research paper is published
Vega-Lite introduced a higher-level grammar for interactive graphics. It showed how encoding rules and composition can express layered and multi-view displays concisely.
2017 Feb
Facebook releases Prophet
Prophet packaged a decomposable additive model of trend, seasonality, and holidays behind a two-column interface aimed at analysts. It made forecasting broadly accessible and, in doing so, moved many forecasts onto a single trend model whose behavior is easy to misread.
2017 Apr
Amazon publishes DeepAR
DeepAR trained an autoregressive recurrent network jointly across many related series and produced probabilistic forecasts. It established the global-model approach that later dominated large-panel problems.
2017 May
AAPOR evaluates the 2016 US election polls
The committee found national popular-vote polls were among the most accurate since 1936, but many state polls underestimated one candidate, partly because they did not weight on education. It made education weighting and late-deciding adjustment standard checks for election polls.
2017 May 17
Firebase Analytics adopts the Google Analytics name
Google unified its app analytics reporting under the Google Analytics brand after the Firebase platform expansion. App behavior and web analytics moved toward one product family.
2017 Jun
Power BI Report Server adds an on-premises publication path
Report Server allowed Power BI reports to be published and viewed within customer-managed infrastructure alongside the Power BI Premium offering.
2017 Aug
Microsoft documents metric-interpretation pitfalls
Microsoft's paper collected twelve recurring mistakes in interpreting online controlled-experiment metrics. It shifted attention from running a test to checking whether the chosen metrics support a trustworthy decision.
2017 Aug
The Dirty Dozen metric pitfalls are published
Microsoft researchers cataloged recurring interpretation errors observed across thousands of experiments. Metric validity and practical interpretation became explicit parts of trustworthy analysis.
2018
Deequ brings declarative checks to Spark pipelines
Deequ made data assumptions executable over large datasets and added automated validation for machine learning pipeline inputs.
2018
DuckDB development begins at CWI
Mark Raasveldt and Hannes Mühleisen start DuckDB at Centrum Wiskunde & Informatica in Amsterdam as an embeddable analytical database project.
2018 Jan
Hyper becomes the Tableau extract engine
Tableau 10.5 introduced Hyper for extract ingest and analytical query processing, moving existing extracts toward the Hyper format during refresh.
2018 Apr
Large-scale experimentation platform anatomy is documented
The architecture separated portal, execution, log processing, and analysis services. This made the end-to-end measurement path a platform design concern rather than only a statistical calculation.
2018 Apr
Tableau Prep becomes available
Tableau Prep added a coordinated visual flow, profile, and data grid for cleaning, combining, and reshaping data before workbook analysis.
2018 Jun
The M4 competition tests machine learning at scale
Across 100,000 mixed series, combinations of statistical methods took most of the top places and the pure machine-learning entries underperformed, while a hybrid of exponential smoothing and a recurrent network won. Combining methods beat any single method.
2018 Jun 5
WCAG 2.1 becomes a W3C Recommendation
WCAG 2.1 added guidance addressing more mobile, low-vision, and cognitive accessibility needs. Dashboard testing expanded beyond a desktop mouse-and-screen workflow.
2018 Aug
Large-scale declarative data verification is demonstrated
A production-oriented system translated reusable data quality constraints into aggregate queries and supported incremental validation.
2018 Sep
Excel previews dynamic arrays
Microsoft previewed dynamic arrays, allowing formulas that return multiple values to spill into neighboring cells.
2018 Sep
TensorFlow Data Validation is introduced
TFDV connected dataset statistics and schemas with continuous validation, anomaly reporting, and training-serving skew detection.
2018 Nov
Composite models become generally available
A semantic model could combine Import and DirectQuery sources, making storage mode a per-model design tradeoff instead of a single global choice.
2019
SIGMOD demo paper published
The peer-reviewed demo paper "DuckDB: An Embeddable Analytical Database" documents the design goals behind the early public release.
2019 May
N-BEATS shows pure deep learning can win
A deep stack of fully connected layers with a built-in decomposition structure beat the M4 winner without statistical components, marking the point where neural architectures became competitive on standard benchmarks.
2019 Jun
DuckDB v0.1 announced at SIGMOD
The first preview release and conference announcement present an analytical SQL engine that runs as a binary or inside host processes such as Python and R.
2019 Jul
Microsoft publishes a sample-ratio-mismatch taxonomy
The paper organized causes of observed assignment counts differing from the configured ratio across assignment, execution, log processing, and analysis. It made sample-ratio mismatch a formal diagnostic before product decisions.
2019 Jul
SRM diagnosis receives a practitioner taxonomy
The KDD paper organized sample ratio mismatch causes across assignment, execution, logging, and analysis. SRM became a concrete validity gate with diagnostic rules.
2019 Jul 31
App plus Web properties are announced
The new property type combined app and website data with a consistent event-based model. It also introduced cross-platform funnels, paths, and ad hoc analysis.
2019 Aug
Microsoft announces XLOOKUP
Microsoft announced XLOOKUP as a successor to VLOOKUP, adding a newer lookup option to the spreadsheet formula toolkit.
2019 Aug
Salesforce completes its Tableau acquisition
Tableau became part of Salesforce, linking its analytics platform to a wider customer and application ecosystem.
2019 Oct
Python 3.8 adds assignment expressions
Python 3.8 added assignment expressions through the colon-equals operator. Expressions can now bind a name where that form improves a condition or comprehension.
2020s
26 milestones
2020
pandas 1.0 marks a project milestone
The 1.0 release documented pandas' maturation into a widely used data-analysis foundation and a stable basis for reproducible EDA work.
2020 Jan
pandas 1.0 establishes its major-version policy
pandas 1.0 formalized a deprecation and major-release policy, making API evolution more legible for long-lived analysis code.
2020 May
Relationships reshape multi-table data modeling
Tableau 2020.2 introduced logical relationships that preserve table detail and combine relevant tables in the context of each view.
2020 Jun
The M5 competition is won by global gradient-boosted trees
On Walmart's hierarchical retail data, every top method was a pure machine-learning model trained across series, and all beat the statistical benchmarks. Calibrated prediction intervals proved harder to produce than accurate point forecasts.
2020 Sep
Deployment pipelines become generally available
Power BI Premium gained managed development, test, and production stages for promoting content while retaining environment-specific connections.
2020 Oct 14
Google Analytics 4 becomes the default for new properties
Google renamed the App plus Web foundation as Google Analytics 4 and made it the default experience for new properties. Event-based cross-platform measurement became the forward product model.
2020 Nov
LET becomes generally available in Excel
Microsoft made LET generally available, letting a formula name and reuse intermediate calculations.
2020 Dec
The UK Government Data Quality Framework is published
The framework organized operational data quality around completeness, uniqueness, consistency, timeliness, validity, accuracy, and user needs.
2021
DuckDB Foundation incorporated
An independent foundation takes custody of most project intellectual property so DuckDB remains MIT-licensed with long-term continuity separate from any single company.
2021 Jul
AAPOR evaluates the 2020 US election polls
Reviewing more than 2,800 polls, the task force reported that 2020 polls overstated one candidate's margin by more than in 2016 and could not identify a single cause, ruling out simple electorate composition assumptions. It renewed attention to differential nonresponse as a hard, unresolved problem.
2021 Jul 14
DuckDB Labs announced
The core engineering team forms DuckDB Labs as a CWI spin-off for full-time development and commercial support around the open-source engine.
2021 Oct
Python 3.10 adds structural pattern matching
Python 3.10 added match and case statements for structural pattern matching. Code can dispatch on the shape of a value as well as ordinary Boolean conditions.
2022 Sep
DuckDB 0.5.0 release line
The 0.5 series continues the named-release cadence as the engine moves through wider production adoption ahead of a stability-focused 1.0.
2022 Oct
Python 3.11 improves tracebacks
Python 3.11 added fine-grained error locations in tracebacks and delivered interpreter speed improvements. Diagnosing an expression failure gained more precise source context.
2022 Nov
Long-term partnership with MotherDuck
DuckDB Labs partners with MotherDuck to provide a managed cloud path for shared storage, collaboration, and operations on top of DuckDB.
2023
ISO publishes the 2023 SQL framework edition
ISO/IEC 9075-1:2023 published the sixth framework edition and references SQL property graph queries as part of the standard family. SQL continues to grow through coordinated parts rather than one vendor's dialect alone.
2023 Jul 1
Standard Universal Analytics stops processing new data
Standard Universal Analytics properties stopped processing new hits. New collection shifted to GA4 properties and their event-based data model.
2023 Oct 5
WCAG 2.2 becomes a W3C Recommendation
WCAG 2.2 added success criteria while retaining backward compatibility with earlier WCAG 2 versions. Report and dashboard teams gained a new current accessibility target for web delivery.
2023 Nov
Microsoft Fabric reaches general availability
Power BI became one workload in a unified analytics platform, with OneLake and Direct Lake connecting report models to Fabric-managed data.
2024
MotherDuck contributes to Node.js client work
Ecosystem collaboration continues as MotherDuck contributes substantially to the DuckDB Node.js client later in 2024.
2024 Feb
Tableau Pulse reaches general availability
Tableau Pulse introduced a metrics-centered experience with personalized visual and natural-language insights in Tableau Cloud.
2024 Mar
Amazon releases the Chronos foundation model
Chronos tokenized series values and trained language-model architectures on a large collection of datasets, producing forecasts for an unseen series with no fitting step and trading control for zero setup.
2024 Jun
NumPy 2.0 is released
NumPy 2.0 marked a major release of the numerical layer beneath many Python analysis workflows.
2024 Jun 3
DuckDB 1.0.0 released
Version 1.0 emphasizes storage backward compatibility and a more cautious SQL and C API change policy so applications can depend on clearer compatibility expectations.
2025 Mar
DuckDB Local UI introduced
DuckLabs and MotherDuck ship a browser-based local UI via the ui extension, with an explicit path to connect to MotherDuck when cloud features are needed.
2025 May
DuckLake format introduced
DuckLabs introduces DuckLake, keeping lakehouse catalog metadata in a SQL database while storing table data as open Parquet.