DataWeaver.app
Publication-ready statistics without SPSS: an R Shiny SaaS with AI-written reports.
- analysis modules
- 39
- report templates
- 18
The Challenge
Researchers and students need publication-ready statistics – reliability, factor analysis, regression, ANOVA – without SPSS and without learning to program in R or Python.
The Approach
I built DataWeaver as a subscription web app in R Shiny. 39 Shiny modules cover data import and cleaning, variable transforms, reliability, exploratory and confirmatory factor analysis, regression, ANOVA and group comparisons, all with APA-formatted tables and AI-written interpretation via the Claude API. Supabase handles auth and the Postgres backend, and Stripe handles monthly, annual and team subscriptions through a Plumber webhook API. The app ships as a Docker image deployed by GitHub Actions to a Hetzner VPS, with Playwright end-to-end tests.
Key Features
- 39 analysis modules: import and cleaning, transforms, reliability, EFA/CFA, regression, ANOVA and group comparisons
- Advanced psychometrics: IRT, differential item functioning (DIF) and measurement invariance
- Mediation and moderation, ML prediction, and qualitative/sentiment analysis
- APA-formatted tables
- AI-written interpretation via the Claude API
- Supabase auth and Postgres backend
- Monthly, annual and team subscriptions through Stripe, via a Plumber webhook API
- 18 report templates, from the analysis straight to a finished report
- White-label deployment for Truity, connected to their live survey data
Results & Impact
- Live SaaS built solo, from concept to production
- White-labelled instance running for Truity
- Automated deploys: Docker image via GitHub Actions to a Hetzner VPS, with Playwright end-to-end tests
- APA-formatted tables and AI-written interpretation via the Claude API