 Today, I'm going to talk about automation of statistics summary and analysis using R-Shine. So what's the motivation for this? Basic statistics analysis, such as summary statistics, regression analysis, and testing of associations are very common in medical research. The researchers need to repeat the general is summary tables and conduct exploratory analysis before they can have a finalized analysis. Many of our collaborators eager to want make sense of their data, but some of them all have less or no experience of our statistical programming experience, which make them very difficult to conduct study analysis. Repeated work can be tedious and inconsistent and time consuming. As more and more journals require reproducibility, it requires tremendous effort of logging and documenting, which can affect the effectiveness of their research. So what's the hurdles research normally come across? First, probably is to find the right packages, which can be very challenging. At the same time, the syntax differences across packages can be also intimidating. For example, one statistics summary tables, there's multiple packages can do it. Table number one, table one, or GT summary. However, all three of them have different formats. It's really, really confusing. At the same time, typo and variable names can cause confusion too, especially those former SAS user will know that variable name in SAS is insensitive. However, R is very case sensitive regarding variable name. Even though that you managed to get the analysis result to find the right file formats to download it to further modification, it's got to become a hurdles. So what's the solutions based on all the experience we had and the hurdles we come across? So one of the idea is to incorporate many commonly used packages into one so that syntax consistent across analysis and reduced and to be consistent. And remove the need to type the data to avoid those typo and case sensitivity. And because we wanna also the version control issues like our package change over time to be resolved. Lastly, we wanna provide output in the right file formats for the convenience of further modifications or editing. So that we created one Shiny app to make take advantage of the intuitive interactive features of Shiny. So as can see here, it's the snapshot of the Shiny app working one. Left Sunhead is the UI which provides the file uploading variable to chosen interface and analysis options. Right-hand side, we provide the output based on the UI entries. Next, I wanna highlight some feature of the Shiny app just because we are targeted to solve the concern on the issues research hurdles. So we wanna have a click and choose UI, no typo is needed. And secondly, we wanna support commonly used our packages in this app. In the meantime, we'll also provide buttons for download the analysis results in different file formats such as Excel and other formats. The interactive exploratory analysis within the app provides a convenient way for the researchers so that they don't need to like back and forth, back and forth and save the effort of doing this kind of work. Because in the Shiny app, the package was locked in, so which make it the reproducible possible because it's the locking issues, locking package versions. Next, I would like to do a quick demo about how to use this app. As you can see, there is some introduction videos included for this researchers who have no experience before to how to use it. And I'll go ahead to do the demo. I'll choose one of the demo files which was HIPAA compliance is a marketing table. Do you wanna make sure it is? And categorical, I would say I wanna generate for maturity insurance information, continuous variable, I wanna last for stay follow updates. Please notice the sequence I click. The group variable, I wanna use treatment. So let's do the histogram to do the exploratory analysis. Look at the distribution of the two variables. And if you click on to compare the results, you see a lot of updates, difference by treatment groups here, clearly here, this is not a file shape and normal distribution doesn't apply. So based on this, we can adjust by this now normal for say follow updates, obviously, right? So we click on the table with few values to see how the result goes. So you can tell that the characteristics was followed by the sequence of the click. And the follow up was calculated using now normal distributions with a node there and a median into quantile. For last day, we're assuming it's normal. So it's a mean standardized was generated. So you can click the itself and to download it. So you can do further modifications. So that's it for table one. If you wanna do analysis, what you can do is, for example, let's do remove one of this and I'm interested in the regressions and I'll click this out. I wanna do expert as the outcome and the two. So because of the binary, I was hoping with logistic regression would be generated. Let's see what's the result look like. So here you can see that logistic regression was spared outcome has been generated. As usual, the sequence was followed the order of your click, universe analysis and multivariate analysis as a ratio was created here. The same for the plots, forest plots with the multivariate analysis result was generated here and just as the table, we can also download this analysis result like CSV files that will be generated over there. So basically that's what can be done using the Shiny app. So let's go back to see that what we want to present here is let you know that the Shiny app can incorporate many variables and many available packages into one which were very useful. So there's some improvements later on for our next versions we're still working on that. I want to thank you for coming to my presentations. I want to thank all our studios, the package developer and Stutter found from CHOP and U-Pen and partial supported by NIH Phonic. We appreciate your comments of feedback. Thank you again.