Developer
In my leisure time, I engage in the hobby of software development.
R Packages
multiverseRCT
multiverseRCT is an R package for conducting multiverse analyses on randomized controlled trial (RCT) data. Instead of making a single set of analytical choices, multiverse analysis explores how different preprocessing decisions and model specifications affect research conclusions.
Key Features:
- Apply multiple preprocessing methods (e.g., different imputation strategies, outlier handling)
- Fit multiple model specifications (adjusted/unadjusted, different model families)
- Visualize the variation in treatment effects across the multiverse
- Quantify the sensitivity of conclusions to analytic decisions
BriDGE
BriDGE is a comprehensive R package for causal analysis of randomized controlled trial (RCT) data. It seamlessly integrates causal discovery through Directed Acyclic Graphs (DAGs) with flexible mediation analysis using Generalized Additive Models (GAMs).
Key Features:
- Causal Discovery: Multiple algorithms (MMHC, HC, PC-stable) to learn causal structures from data
- Flexible Mediation Analysis: Handles nonlinear relationships using GAMs with automatic smoothing
- Bootstrap Inference: Robust uncertainty quantification through parallel bootstrapping
- Sensitivity Analysis: Assess robustness of findings to data perturbations
- Comprehensive Visualization: Publication-ready plots for all analysis components
- Parallel Processing: Efficient computation for large-scale analyses
- Robust Error Handling: Graceful handling of convergence issues and edge cases
sequenceRCT
sequenceRCT is an R package for analyzing behavioral trajectories in randomized controlled trials. It shifts focus from endpoint-only comparisons to tracking how participants transition across discrete states over time, making it particularly suited for behavioral public policy research where timing, persistence, and relapse patterns are critical to intervention design.
Key Features:
- State Encoding: Converts raw state labels into coded sequence matrices
- Transition Analysis: Estimates empirical transition probabilities between adjacent time points
- Complexity Profiling: Computes entropy, turbulence, and volatility metrics for participant sequences
- Trajectory Clustering: Groups similar behavioral patterns using TraMineR distances and hierarchical clustering
- Statistical Testing: Performs inferential tests for between-group differences in complexity measures and cluster membership (Wilcoxon, Cliff’s delta, Chi-square)
- Visualization: Generates publication-ready plots for states, transitions, complexity, and clusters
The main entry point is analyze_rct_sequences(), which executes the complete pipeline and returns comprehensive results.
macOS Apps
ArticleOrbit
Inspired by behavioural science principles and built by an academic who couldn’t find the right tracker, ArticleOrbit is the must-have macOS app for managing every manuscript milestone—effortlessly and intuitively.
Why ArticleOrbit?
- Behaviour-Driven Design: Interfaces and reminders shaped by proven behavioural science
- Unified Submission Tracker: Log titles, venues, authors, and all key dates
- Live Dashboard & Analytics: Status breakdown, journal stats, waiting times
- Timeline View: Stacked bars map your submission cadence over time
- Quick Sharing & Export: Share summaries via Mail or Notes; export to CSV
- Flexible Search & Sort: Filter by title, journal, or status instantly
- Secure Attachments: Link manuscript files with macOS sandbox security
Perfect for: Principal investigators, research groups, and individual scholars who value clarity, data, and deadlines.
iOS Apps
Randomista
Currently in development. More information coming soon.