Skip to content

GCRI · Data and Compute

Reproducible pipelines

A result can be trusted when others can rerun the same analysis on the same data and get the same answer, and when they can trace each number to its source. Reproducible pipelines combine versioned data and code, recorded environments, workflow systems and provenance. This page sets out why analyses fail to reproduce, how reproducibility is engineered and checked, and what evidence serves the researchers, publishers and agencies that depend on it.

01Ilaphu lokuzoba

Indlela umsebenzi oqhuba ngayo.

Iinkqubo · ezijikeleze esi sifundoNgumzekelo

Data platform and pipelines

Reproducible pipelines · Public data platform · Catalogue v9

Rights review openGCRI · Data and Compute

Datasets

Flood extentsVulaCompleteEwePublished
Health facility listVulaCompleteEwePublished
Mobility tracesRestrictedPartialayisebenziRights review
Household surveyRestrictedCompleteEweAccess by request

Pipeline checks

Pinned dependencies, %76 of 100 requiredDeterministic outputs, %90 of 95 requiredProvenance recorded, %95 of 100 required
Data platform and pipelines: the screen's structure, terms and states for Reproducible pipelines.Example names and illustrative values; a design, not a record or a live system.

Fig.Data platform and pipelines: example names and illustrative values, not a record or a live system.

02The question

Reproducible pipelines: how the risk develops.

Pipelines, models and experiment provenance. The dossier below sets out how the risk develops, what is measured, where the data come from, how it is modelled and who decides.

Indlela yokusebenza

Environments drift

Software libraries, operating systems and hardware change. An analysis that ran last year may give different results, or none, without a recorded environment.

Indlela yokusebenza

Data versions change

Datasets are corrected and updated. Unless the exact version used is recorded, results cannot be traced or rerun.

Indlela yokusebenza

Manual steps break chains

Steps done by hand in spreadsheets or interactive sessions are not recorded and cannot be repeated exactly.

Indlela yokusebenza

Provenance links results to sources

Recording which inputs, code and parameters produced each output lets errors be traced and corrections propagate.

03Ijografi

Connected activity in the contiguous United States.

Night lights, population density and true-colour imagery over the contiguous United States, from NASA GIBS. Each layer carries its own date.

Imephu esebenzisanayo idinga i-JavaScript; uluhlu lweeleya nemithombo zingezantsi.

Iileya 3

Connected activity, for this page

Okutsha okukhoyo

NASA GIBS / EOSDIS · ASTER GDEM (METI, NASA) · Blue Marble (NASA) · © abanegalelo kwi-OpenStreetMap (ODbL) · Natural Earth

Kulayishwa imifanekiso yamva nje ye-NASA.

Rhuqa ukuze uhambise. Ctrl okanye ⌘ uze ujikelezise, chophela okanye usebenzise amaqhosha ukusondeza; beka isalathisi phezu kwento ukuze ubone amaxabiso aphantsi kwaso; ukubuyisela kubuyela kummandla.

Connected activity, framed on the contiguous United States.Satellite and model products as NASA publishes them, each with its own date. They are indicators, not assessments of any place.

Data: NASA GIBS · ASTER GDEM · Blue Marble · © OpenStreetMap contributors · Natural Earth · NASA EONET · NASA HLS S30Sources and licences

  • We acknowledge the use of imagery provided by services from NASA's Global Imagery Browse Services (GIBS), part of NASA's Earth Science Data and Information System (ESDIS). Colours redrawn from the published colour map; values averaged into hexagonal cells where shown (modified). NASA open data
  • We acknowledge the use of imagery provided by services from NASA's Global Imagery Browse Services (GIBS), part of NASA's Earth Science Data and Information System (ESDIS). Colours redrawn from the published colour map; values averaged into hexagonal cells where shown (modified). NASA open data
  • VIIRS Black Marble night-time lights, NASA Goddard Space Flight Center; distributed by the NASA Level-1 and Atmosphere Archive and Distribution System (LAADS DAAC); imagery via NASA GIBS. Colours redrawn from the published colour map; values averaged into hexagonal cells where shown (modified). NASA open data
  • Center for International Earth Science Information Network (CIESIN), Columbia University. 2018. Gridded Population of the World, Version 4 (GPWv4): Population Density, Revision 11. NASA Socioeconomic Data and Applications Center (SEDAC). Licensed under CC BY 4.0. Colours redrawn from the published colour map; values averaged into hexagonal cells where shown (modified). doi:10.7927/H49C6VHW. CC BY 4.0
  • VIIRS near-real-time products (flood, land surface temperature, snow cover, corrected reflectance), NASA LANCE; imagery via NASA GIBS. Colours redrawn from the published colour map; values averaged into hexagonal cells where shown (modified). NASA open data
  • MODIS land products (vegetation indices, gross primary production, land cover), NASA LP DAAC and LANCE; imagery via NASA GIBS. Colours redrawn from the published colour map; values averaged into hexagonal cells where shown (modified). NASA open data
  • ASTER GDEM is a product of METI and NASA (ASTER Global Digital Elevation Model V003, NASA LP DAAC); imagery via NASA GIBS. Colours redrawn from the published colour map; values averaged into hexagonal cells where shown (modified). Cite as: NASA/METI/AIST/Japan Spacesystems and U.S./Japan ASTER Science Team (2019). ASTER Global Digital Elevation Model V003. doi:10.5067/ASTER/ASTGTM.003. NASA open data
  • Yenziwe nge-Natural Earth. Idatha yemephu ye-vector ne-raster esimahla @ naturalearthdata.com. Yeyoluntu
  • © OpenStreetMap contributors (ODbL), via NASA GIBS (coastlines, boundaries, roads and place names). Colours redrawn from the published colour map; values averaged into hexagonal cells where shown (modified). ODbL
  • Blue Marble, NASA Earth Observatory; imifanekiso ivela kwi-NASA GIBS. Imibala izotywe ngokutsha ukusuka kwimephu yemibala epapashiweyo; amaxabiso afakwe umndilili kwiiseli ezineheksagoni apho aboniswa khona (iguqulwe). Idatha evulekileyo ye-NASA
  • Imeko kwigloubhu nakwimephu ethe tyaba: unxweme, imida, iindlela namagama eendawo avela kwiileya zesalathiso ze-NASA GIBS (idatha ye-OpenStreetMap, ODbL, © abanegalelo kwi-OpenStreetMap); imifanekiso ye-Blue Marble njengesiseko esikhethwayo; umhlaba we-3D nokufundwa kokuphakama okuvela kwisalathiso semibala sokuphakama se-ASTER GDEM (iiklasi ezisuka kwi-5 m ukuya kwi-200 m, zandisiwe ukuze zifundeke). Imida izotywe njengoko ipapashiwe kwaye ayibonisi kuma kuwo nawuphi na ummandla.
  • Iziganeko ezenzeka ngoku: NASA Earth Observatory Natural Event Tracker (EONET); iziganeko zingumqondiso nje kwaye zigcina amakhonkco azo emithombo yokuqala.
  • Observed (EO): NASA HLS S30 (contains modified Copernicus Sentinel data 2026). We acknowledge the use of imagery provided by services from NASA's Global Imagery Browse Services (GIBS), part of NASA's Earth Science Data and Information System (ESDIS).
  • NASA GIBS: Siyavuma ukusebenzisa imifanekiso enikwe ziinkonzo ze-NASA's Global Imagery Browse Services (GIBS), eyinxalenye ye-NASA's Earth Science Data and Information System (ESDIS).
  • Iileya ezikhethwayo (imithombo emi-5 kuluhlu lweeleya: ephilayo, uqikelelo, eqwalaselweyo nemeko) zibonisa isincomo somboneleli wazo kwimephu nalapha xa zivuliwe; uqikelelo sisiphumo semodeli kwaye izalathisi ezivelisiweyo ziqikelelo.

Iileya nemihla

  • Night-time lights, daily: VIIRS Black Marble (VNP46A2), NASA LAADS DAAC · 12 Oct 2026
  • Population density, 2020: Population: CIESIN, Columbia University, Gridded Population of the World v4.11 (2018), NASA SEDAC, CC BY 4.0 · static
  • True colour: VIIRS NOAA-20 corrected reflectance, NASA LANCE · 13 Oct 2026
  • Uhlobo lwesigubungelo somhlaba (IGBP), ngonyaka: MODIS Terra and Aqua isigubungelo somhlaba (MCD12Q1), NASA LP DAAC · 2024
  • Enhanced vegetation index, rolling 8 days: MODIS Terra EVI, NASA LANCE · 13 Oct 2026
  • Vegetation greenness (NDVI), rolling 8 days: MODIS Terra NDVI, NASA LANCE · 13 Oct 2026

04Across the domains

Reproducible pipelines let each domain's results be checked.

Pipeline methods are designed for the Methods and Technical series. FAIR principles for research software (2022) and PROV-O (2013) guide practice; lineage would follow NS-3b of the proposed Nexus Standards.

  • Reproducibility audits.
  • Pipeline templates with tests.
05Isilinganisi sokuphindwa nothotho lweziphumo

Model reproducible pipelines as a system. Run the cascade.

Assembled from the Nexus Standards area profile Digital, telecom and platforms: the hazards it names, the systems and sectors that carry the loss, the controls that dampen it and the data that observe it. Choose the components, pick an initiating hazard and step through the cascade. The Map view places the run on an example geography with open data. Every coupling weight is a stated assumption you can change. A teaching model, not a forecast, an assessment or a certification. To model a place of your own choosing, open the full System Builder.

Iingozi

Izilawuli

  • NS-0 Umgaqo-siseko nemigaqo
  • NS-9 Ii-UNI zemeko (ukusasazwa, iimeko, iimeko zenkqubo)
  • NS-11 Ukuqinisekiswa nokubekwa kweliso (ubungqina, iziqinisekiso, uqwalaselo lokuhambelana ngaphandle kokuthemba)
  • NS-12 Software, model and AI/ML artefacts (supply chain, model cards, dataset sheets)
  • NRS-H86 Hazard family: Telecom and data-centre failure (NHS-1 infra.digital)
  • NRS-H54 Hazard family: Signal and platform disruption (NHS-1 cyber.platform)
  • NRS-H93 Hazard family: Identity, trust and cryptographic failure (NHS-1 cyber.identity)
  • NRS-H94 Hazard family: Systemic cyber events (NHS-1 cyber.systemic-cyber)
  • NRS-T8 Software and supply-chain integrity risk
  • NRS-X6 Cyber and digital security risk

Uqwalaselo

  • ITU DataHub ICT indicators (listed for reference)
  • World Development Indicators (listed for reference)
  • WMO Alert Hub and Register of Alerting Authorities (listed for reference)
  • Proportion of population covered by a mobile network, by technology
  • Fixed broadband subscriptions per 100 inhabitants, by speed
  • Individuals using the Internet (% of population)
  • Number of people per 100,000 that are covered by early warning information through local governments or through national dissemination mechanisms

Umzekelo wejografi

  • North America: 243 hexagonal cells
  • Izibane zasebusuku namhlanje, ukuqhawuka kombane (VIIRS Day/Night Band), NASA GIBS
  • Vegetation disturbance alerts, 30 m (OPERA DIST-ALERT), NASA GIBS
06Izalathisi

What is measured, and how well.

Each indicator carries its unit, its cadence, the kind of source behind it and the evidence label a value would carry. No values are shown: a programme supplies them under its own authority.

IndicatorUnitCadenceSource typeIleyibhile
Published analyses with code and data availablepercent of outputsper auditPublication auditsKUQWALASELWE
Analyses rerun successfully by a third partypercent of attempted rerunsper auditReproduction studiesKUQWALASELWE
Datasets with persistent identifierspercent of datasetsannualRepository metadataKUQWALASELWE
Pipelines with recorded environmentspercent of production pipelinesper assessmentPipeline repositoriesDECLARED
Corrections issued after reproduction checkscountannualCorrection logsKUQWALASELWE
Software archived with identifierscountannualSoftware archive recordsKUQWALASELWE
07Methods and evidence

How reproducible pipelines is modelled, measured and tested.

The model, its assumptions, its data and its uncertainty, in the form a scientist or engineer would check. Structures follow published methods. The values drawn are illustrative, not findings.

Datasets and provenance

The data classes behind reproducible pipelines and what is recorded about each before it is used.

Data classProvenance recordedResolution statedLicence checked
Satellite imagery, optical and radarsensor, processing level, acquisition datepixel size and revisitper scene
Reanalysis and climate model outputmodel, experiment, member, versiongrid and time stepper release
In-situ sensors and gaugesstation, instrument, calibration recordpoint and sampling rateper network
Population and asset exposuremethod, vintage, input sourcescell size and unitper release
Boundaries and reference geographyproducer, edition, datescaleper edition
Imvelaphi
W3C PROV record from source to figure
Imigaqo
FAIR: findable, accessible, interoperable, reusable
Gaps
shown as missing, never filled silently
Data classes and the provenance fields recorded for each; a dataset is listed by name only once its licence admits the use.
FAIR guiding principles · GO FAIR and FORCE11W3C PROV · World Wide Web ConsortiumGuidance on consistent treatment of uncertainties · IPCCCEOS Analysis Ready Data · Committee on Earth Observation Satellites
08Imithombo

Where the evidence comes from.

Named public datasets and institutional records, each with its licence class. Sources by agreement are used only under their terms and never republished.

UmthomboWhat it givesLicence class
DataCite and Crossref metadataPersistent identifiers for data, software and publicationsOpen, CC0 metadata
Software Heritage archiveArchived source code with intrinsic identifiersOpen, archive terms
Zenodo and institutional repositoriesDeposited data and code with identifiersOpen, licence per record
W3C PROV standardA model for representing provenanceOpen standard, W3C terms
Reproducibility studies in published literatureMeasured reproduction rates in fieldsLicence per publication
09Decisions

Who decides, and what the evidence serves.

Data and Compute prepares evidence. The decisions below are taken by the authorities named, under their own rules, and never by the platform.

Who decidesThe questionEvidence this page serves
Research teams and analystsHow to structure pipelines so results can be rerunChecklists; pipeline templates
Publishers and report editorsWhat reproducibility evidence to requireAudit methods
Funders and agenciesWhether to fund archiving and reproducibility supportReproduction rate evidence

Outputs this topic produces

  • Reproducibility checklists for analyses and reports
  • Pipeline templates with versioned data, code and environments
  • Provenance recording designs
  • Audit methods for reproduction of published results
10Core questions

What experts ask about reproducibility.

  • Can this result be regenerated from recorded inputs, or does it depend on undocumented steps?
  • Which exact version of the data, code and model produced it?
  • If an error is found upstream, which downstream results are affected?
  • Will the pipeline still run in years, after software and hardware have changed?
  • Can a reviewer without the original team understand what each step does and why?
  • Are random seeds, model weights and configuration choices recorded with each result?
11Mechanisms

Why results become irreproducible.

Irreproducibility is rarely deliberate. It accumulates through small, reasonable shortcuts made under deadline pressure, each harmless alone. The cost appears later, when a result is questioned, a correction is needed or a team member leaves.

Environments

Software drift

Libraries and dependencies change, so the same code produces different results or stops running.

Idatha

Moving inputs

Datasets are updated in place without versioning, so the original input cannot be recovered.

Inkqubo yokusebenza

Manual steps

Undocumented manual cleaning or parameter choices cannot be repeated by others.

12Methods and limits

How pipelines are made reproducible.

  1. 01

    Version control

    Tracking code, configuration and, where possible, data versions. Limits: large datasets need specialised tools.

  2. 02

    Containerised environments

    Packaging software dependencies so they can be rebuilt. Limits: containers themselves depend on underlying systems that change.

  3. 03

    Workflow systems and provenance

    Tools that run pipelines and record each step, studied under data engineering, interoperability and stewardship. Limits: adoption requires effort and discipline.

  4. 04

    Numerical reproducibility

    On parallel and accelerated hardware, floating-point results can differ slightly between runs, a matter for computing specialists. Limits: exact bitwise reproducibility is sometimes impractical, so tolerances must be stated.

13Iindlela zokusebenza

How it is studied, and what the methods cannot say.

14Where this page's references

Apho izalathiso zeli phepha zikhokelela khona.

Amaqumrhu namaxwebhu eli phepha elifundwa ngokuthelekiswa nawo. Imeko neenguqulelo ziyatshintsha; ikhonkco ngalinye liya kwiqumrhu elikhuphayo.

15Imibuzo

Imibuzo neempendulo.

What is the difference between reproducibility and replicability?
Reproducibility usually means getting the same results from the same data and code; replicability means getting consistent results from new data, though usage varies by field.
What is a persistent identifier?
A long-lasting reference to a digital object, such as a DOI, that continues to resolve even if the object's location changes.
What is data provenance?
A record of the origin of data and every transformation applied to it on the way to a result.
Is reproducibility the same as correctness?
No. A wrong analysis can be perfectly reproducible. Reproducibility makes errors findable and correctable.
Why can results differ on different hardware?
Parallel computations can add numbers in different orders, and floating-point arithmetic is not perfectly associative, so tiny differences can arise.
Does the platform review other teams' pipelines?
No. It publishes patterns that teams can apply themselves.
IingxeloNikela

Faka ubungqina obuvela kweli phepha.

Reports is the Consortium's quarterly publication. What you know about this subject can become evidence, a case or a correction in the next edition, credited to you and published under CC BY 4.0 where it is your own work.

GCRIData and ComputeExponential TechnologyStrategic Systems
Data and Compute

Bring a method, a dataset or a question.

Information and enquiries are open today. An enquiry receives a reference and names the next step; applications, payments, compute allocations and registrations open only when their team, terms and review capacity exist.