Week 7

Here are my tasks for week 7:

Task: Data visualization of expression data and user customization of graphic parameters such as color palettes, labeling, model fitting, multiple graphs and interactivity.

Method: Develop boxplots of expression data and other informative graphics using R packages: ggplot2, pheatmap, htmlwidgets, RColorBrewer. Generate an interactive components that allow a user to visualize different statistical variables on the graphics using the rCharts R package. The use of JavaScript visualization libraries will incorporate flexibility for the user and make the best use of descriptive graphics suitable for both on-line and offline publication of results.

---

July 7

• I merged two pull requests from Dr. Dancik that improves: Correctly resets clinical data when user changes platform; uses alert if gene symbol not found, and Make sure selectGene reactive returns NULL if gene column is not found.

July 9

• I changed parse.modal dataframe to a reactive function that returns the selected columns as a summary similar to the clinical data summary page - this change makes viewing all of the selected parameters for time, outcome, and x easier for editing and viewing

• I'm currently looking into updating the base graphics for the three plots (Boxplot, stripchart, and Kaplan-Meier curve) that are currently on the application. I've found the first way to Creating good looking survival curves – the 'ggsurv' function via this neat tutorial article which uses the ggsurv function in the GGally R package. Ultimately I think the format of using ggplot2 themes like this is the cleanest way to get modern looking graphics that users will be able to utilize outside of the application into publications or other reproducible reports.