svyLab’s cover photo
svyLab

svyLab

Data Infrastructure and Analytics

Hoboken, NJ 484 followers

Survey design and analysis you can defend. Built on the open-source svy engine.

Follow

About us

svyLab turns a questionnaire and an analysis plan into weights, estimates, and a branded report, with the method behind every number visible to a reviewer. Most survey teams keep the sampling design in one document, the weights in a script, and the forty standard tables in another script that one person fully understands, with the final numbers typed into a spreadsheet by hand. When a reviewer asks which weight, which variance method, and where a margin of error came from, reconstructing the answer takes days. When that analyst leaves, it takes longer. svyLab closes that gap. The analysis plan is written once and the methodology acknowledged once by a person, and every wave after that is: attach the new file, preflight, run, report. Each step is versioned, checked before it runs, and restated in the terms a methods appendix would use. It is built on svy, the open-source Python engine for complex survey design, weighting, and estimation, validated against R and open source since 2020. It is made for national statistical offices and ministries, for public health and development survey programs including DHS, MICS and LSMS and their implementers, and for survey research firms and university survey centers. Anyone who re-runs the same tables every round from a script only one person fully understands. svyLab is onboarding five design partners ahead of general availability. It starts with a free 30-minute walkthrough at svylab.com/demo, on a public survey or on your own published tables, with nothing to install and no data to send. svyLab™ is a product of Samplics LLC.

Website
https://svylab.com/demo?utm_source=linkedin&utm_medium=social&utm_campaign=company-page
Industry
Data Infrastructure and Analytics
Company size
2-10 employees
Headquarters
Hoboken, NJ
Type
Privately Held
Founded
2022
Specialties
survey methodology, complex survey design, survey weighting, raking and calibration, non-response adjustment, variance estimation, design-based estimation, small area estimation, official statistics, household surveys, DHS, MICS, LSMS, questionnaire design, sampling, survey data analysis, reproducible research, Python, statistical software

Locations

Updates

  • View organization page for svyLab
    svyLab

    484 followers

    The World Bank made reproducibility packages standard for its research. svyLab makes it a by-product of the work.

    View profile for Mamadou S Diallo
    Mamadou S Diallo

    The World Bank now publishes a reproducibility package with every research finding it releases. Will survey programs be next? Some context. According to a Nature editorial this month, since 2025 the World Bank has accompanied all the research findings it publishes with reproducibility packages (data, code, and documentation of how the numbers were produced), with more than 500 released and an internal team verifying packages for a badge. The Inter-American Development Bank is following. The editorial calls on UN agencies and other international organizations to do the same. With svyLab, reproducibility is by design: data, design, weights, code, and results travel together, so reproducing an estimate is a rerun, not a reconstruction. The reproducibility package is a by-product of doing the work, not a project after it. If your program publishes survey estimates and the reproducibility question is starting to arrive, I'd be glad to show you what it looks like when the answer is already there. DM me, or book a walkthrough at svylab.com. #OfficialStatistics #Reproducibility #SurveyResearch #DataGovernance

Similar pages