Eshwar Udho, PhD

Skills: R Programming. Data Manipulation. Exploratory Data Analysis. Statistical Analysis and Modeling. Publication-Quality Figures. Teaching. Research. Experimental Design. My Data Science skills have come about from a pseudo-random walk down the open-source statistical software lane in the physical sciences - Initially I learned R (Udemy) to not have to learn new statistical software with each new lab I entered. While for some time Excel array formulas seemed like the best thing in the world, R soon replaced Excel almost entirely, both for compiling and formatting data, as well as for making figures, especially publication-quality figures. At my current job, my boss knows I am pretty good with R, so I usually get things like, "Can you make me a figure to show variation in this data if my main point is..." or, after seeing one of my figures, "our collaborator did this LDA and Clustering Analysis, it looks like this, can you do that too" (I did not, but Google is an awesome thing). It's fun with a boss like that, I can now rattle off the packages that will be in every one of my R scripts or Notebooks and my grasp of many concepts are better than ever since I have to explain them to my boss (ROC curves, cross-validation, ... ) or explain to my boss that I know them (how does he know about bootstrapping?!). My R skills were acquired by learning on-the-fly as we in the lab needed to do something new. To get some formal training, I have taken 2 Udacity nanodegrees and between those projects, my daily work and some Kaggle projects, I can now run exploratory data analysis and perform hypothesis testing fairly quickly. I find the most time-consuming step to be compiling and cleaning data, while the most critical step is asking the right questions to make a coherent story of the data. Man, am I thankful for all those computer science classes I took way back in undergrad!