![]() ![]() I have also formed some early impressions and will share some of these thoughts as well. In short, I learned a ton and thought I’d share what I discovered along the way while it is fresh in my mind. This is the story of why, what, and how I did this. I decided to finally build my first R package a few weeks ago. However, I had moved on to another role before we got around to exploring this option. I knew we might benefit from developing R packages to better formalize, govern and promote reuse, even if for internal use. We wrote a lot of R code: mostly scripts to be used interactively by our data scientists or called by other scripts/programs. Most of what we did in R was specific to applications for our company and for internal use. However, I always tried to stay close to the design and implementation details, including what we were doing with R. For me personally, I was in truth spending most of my time on management tasks: recruiting and developing the team, selling the team’s services, and creating a strategy with company leadership and my peers to demonstrate the value of analytics, and grow the capability in the organization. As others have noted, this stuff really is over 80% of the work in the analytics value chain. We also did a lot of data “infrastructure” work: data ingestion, data & metadata modeling, data warehousing, data quality and other things that have come to be identified as data engineering. ![]() In general, we thought about how R might fit into production workflows and services. We prototyped exposing R models as web services for consumption by various clients. We ended up applying R to do a lot of cool things like time series forecasting (great for workforce optimization in a shared services organization), leads generation, and predictive modeling. In fact, our team helped establish the Philippines R User Group which has grown quite a bit since the first meeting in 2013. Our data scientists adapted R for a few reasons including price, the extent of package support and the size of the community. It was a small team and in many ways we were breaking new ground. I was setting up a team at our shared service center to provide analytics services.
0 Comments
Leave a Reply. |
Details
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |