Nicholas CoxSome Graphical Tips for Stata UsersThe talk will cover a miscellany of graphical tips, some old, some new. Nick will discuss using both official commands and community-contributed commands. He will range from small stuff (often it is a matter of detail to change a good graph into a much better one), through various techniques and tricks, to broad strategy, both in learning and using Stata easily and effectively and in working with graphics for research, teaching, and service.Click here for presentation slides Watch presentation |
Aramayis DallakyanIntroduction to Explainable Machine Learning Using StataMachine learning (ML) has become a powerful tool for modeling complex data and providing accurate predictions. However, the "black-box" nature of many ML models often raises concerns about their explainability and trustworthiness. Explainable machine learning (XML) seeks to address these concerns by enhancing the transparency andunderstanding of ML predictions. This talk aims to provide a practical guide to XML techniques. It begins with an overview of ensemble decision tree models such as random forests and gradient boosting, which are widely used but often difficult to interpret. Aramayis then introduces methods for explaining predictions using both global and local XML techniques. These include state-of-the-art approaches such as SHAP values, individual conditional expectation (ICE) plots, variable importance measures, partial dependence plots, and global surrogate models.Click here for presentation slides Watch presentation |
Zixuan CongPreliminary Findings on Advancing Women's Health in Singapore through AI AcceptanceAs artificial intelligence rapidly advances, it can integrate electronic health records, genetic profiles, and clinical information to enable individualized, female-specific prevention strategies. This presentation uses structural equation modelling in Stata to analyses survey data to examine factors influencing Singapore women’s adoption of AI-enabled healthcare.Click here for presentation slides Watch presentation |
Pablo GluzmannSAMREGC: Stata Module to Perform Sensitivity Analysis of Main Regression CoefficientsThis presentation introduces samregc, a fast, flexible, and simple Stata command that systematizes specification-based sensitivity analysis. It evaluates the robustness of target coefficients by analyzing all–subsets (or user defined subsets) regression results over alternative combinations of control variables.Click here for presentation slides Watch presentation |
Andrew Gray - Withdrew due to illness"Can I just use n=30 in each group?": Using Stata for Sample Size Determination in an Increasingly Complex WorldIn this talk, Andrew will outline some of the practical and technical challenges involving sample size that he faces as a biostatistician in the health sciences. He will describe how Stata helps to support a workflow including simulations when needed and use examples from recent research projects. |
Yuke LiHow Cooking and Eating at Home Shape Emotional Well-Being: Insights from the Food & You SurveyUsing data from the Food & You Survey, this study aimed to evaluate the behavioural pathways linking emotional well-being, cooking behaviour, and eating-at home practices, and to identify leverage points for public-health and behavioural interventions. Partial Least Squares Structural Equation Modelling was performed with Stata to assess these directional relationships.Click here for presentation slides |
Dean McKenzieHealthcare Quality Control and Improvement Using StataStata is widely used in healthcare to compare events such as falls, infections and episodes of delirium over time using control charts across hospital wards or across hospitals. This presentation demonstrates methods including control charts, funnel plots, ANOM and contrast and user written CHAID with the goal of developing healthcare quality improvement techniques.Click here for presentation slides Watch presentation |
Irma Mooi-Reci & Tim LiaoXTVFREG: Stata Module for Estimating Variance Function Panel RegressionThis presentation introduces xtvfreg, a new Stata module that implements an iterative mean variance panel regression estimator in which both the conditional mean and conditional variance of the dependent variable are modeled as functions of covariates. The estimator is designed for researchers working with panel data in which heteroskedasticity is substantively meaningful.Click here for presentation slides Watch presentation |
Marianna Nittirdlasso: Regression Discontinuity with High-Dimensional DataThis presentation discusses a command, rdlasso, which allows the inclusion of high dimensional covariates in Regression Discontinuity Design (RDD) settings. The command allows for the inclusion of high-dimensional covariates in RDD for sharp and fuzzy cases, making the methodology accessible to Stata users and also automating the covariate selection procedure.Click here for presentation slides Watch presentation |
Mathew PiercyFluid Balance in Postoperative Patients and Relationship to Acute Kidney InjuryThis paper presents an audit of 330 postoperative patients examining the relationship between postoperative cumulative fluid balance over 7 days and the incidence and rate of recovery of acute kidney injury. The data was analysed with Stata using a zero-inflation Poisson regression model and compared to a GEE model and 2 models using h2o machine learning.Click here for presentation slides Watch presentation |
Alannah RudkinThe Use of Stata putdocx for Automating Data Safety Monitoring Committee ReportsData safety monitoring committee meetings for clinical trials necessitate the creation of a statistical report from which the trial's safety, progress, and data integrity can be assessed. The presentation shows how much of the repetitive process can be streamlined and automated using Stata’s putdocx commands via a do-file.Click here for presentation slides Watch presentation |
Thomas SosecoHousehold Net Wealth Inequality in Indonesia: Evidence from a Dagum Type III ModelInvestigating household net wealth inequality in Indonesia is important as it can worsen for low-class individuals or households who are unable to inherit sufficient capital for the next generations and maintain financial stability during a period of low or no income. This paper applies the Dagum Type III model to measure household net wealth inequality in Indonesia.Click here for presentation slides |
Xuelu Sungofbinreg: Goodness-of-fit Statistics in Binary Regression ModelsWhen reporting the results of binary regression, it’s crucial to evaluate the overall model adequacy using goodness-of-fit statistics. This presentation introduces a command gofbinreg, which assesses the performance of the Hosmer-Lemeshow, normalised unweighted sum of squares and Hjort–Hosmer statistics to evaluate overall model adequacy.Click here for presentation slides Watch presentation |
Luyang XiaoYoung Hearts at Risk: Preliminary Insights into Detection and Personalized Management of Acute Myocardial InfarctionThis study investigates acute myocardial infarction among younger adults in Singapore and characterizes their clinical and risk profiles to guide early detection and targeted management. Using data from the National Registry of Diseases Office, a retrospective cohort of patients was analysed for 1-year all-cause mortality using Bayesian proportional hazards models in Stata.Click here for presentation slides Watch presentation |
Ricardo Rodolfo Retamoza YocupicioLimitations and Comparison of the DFA PP and KPSS Unit Root Test: Evidence for Labour Variables of MexicoUnit root tests have represented a great contribution to time series analysis by detecting variable stationarity. However, this presentation includes some of the criticisms that have been made to the unit root tests by executing in Stata the three best-known unit root tests for the main macroeconomic variables of Mexico, this with the intention of analyzing, both graphically and technically, whether the series are stationary or not.Click here for presentation slides Watch presentation |
Hengni YuanArtificial Intelligence in Suicide Prevention: Comparative Evidence from a Network Meta AnalysisArtificial intelligence is emerging as a powerful tool in suicide prevention. Often outperforming traditional assessments, machine-learning models can analyse electronic health records and social-media language to identify subtle behavioural cues that precede suicidal thoughts or actions. This study applies network meta-analysis to the systematic review by Lejeune et al. (2022), which highlights the potential of AI in improving suicide-risk detection, screening, and monitoring.Click here for presentation slides Watch presentation |
Shufan ZhaoMultistate Survival Modelling of Cardiovascular Admission and Mortality in a Heart Failure Cohort in SingaporeHeart failure patients often experience complex clinical trajectories involving hospitalisation and death. Conventional survival models that focus on a single endpoint may fail to capture these sequential outcomes adequately. This study applies a multistate survival framework to characterise transitions to cardiovascular admission and death.Click here for presentation slides Watch presentation |
|
Swift Stata Stories |
|
Mark ChatfieldFinding Incorrect References to Variable Names or Event Names in a REDCap DatabaseManually checking a REDCap database before it goes into production mode is a laborious task. By importing the data dictionary into Stata and then extracting references to [variable-names] and [event-names] in calculations, branching logic etc., I show how some basic checks can be done quickly.Click here for the .do file Watch presentation |
Asmamaw Demis BizunehUnsupervised Clustering in Stata: A Practical Clinical Data ApplicationCluster analysis enables the identification of natural groupings based on similarity across input features. This story demonstrates the key steps involved using Stata commands in combination with R scripts. The story highlights Stata’s fast and transparent workflow for unsupervised clustering.Watch presentation |
Cindy HanUncovering Mathematical Values: A Stata-Powered Bilingual Text Analysis ToolThis story presents the development and application of the Values Automatic Sorting Algorithm, a custom text-analysis tool built entirely within Stata to efficiently process large-scale, open-ended bilingual survey data.Watch presentation |
Ruben Martin-ClarkPlotting Sports Data in StataThis story shows you how to plot several different sports fields, including field hockey, soccer, basketball and others, and then plot player positions on the field for game.Watch presentation |
Gabor MihalaA Minimalist Approach to Version Control and Reproducibility in StataThis story introduces a deliberately, almost too-simple approach to version control and reproducibility for in-house Stata workflows: a short paragraph of Stata code that can be pasted into any do-file. This code automatically records the development history of the file and supports reproducible results.Watch presentation |
|