Overall Satisfaction with RStudio
RStudio is the main tool we use in advanced analytics for predictive modeling development, maintenance, and deployment into production. We run it on a separate Linux server that connects to our enterprise data warehouse. The business problem it addresses is risk modeling on our patient population for outreach and operational efficiency purposes. In addition we use it for data visualization exercises in GGPLOT2 which are more novel than simple excel plots so is a favorite among my team.
- Packages are really easy to install and load.
- Models are really easy to deploy in this language. The functions are simple and relatively straightforward which democratizes the process.
- Links very well with the Oracle data warehouse both to read data and write tables in the EDW.
- Errors in R and RStudio are almost impossible to interpret.
- Updating R and RStudio can lead to packages no longer working which can be frustrating since I have to reinstall all of them from scratch.
- The visual interface is not as good as Jupyter notebook it's tough to read the font especially when projecting onto a screen with projector for meetings.
- We've been able to implement several predictive models that stratify risk among our patient population which has been a positive impact.
- We've conducted several clinical program evaluations using the MatchIt package in RStudio which demonstrated the clinical effectiveness of our programs.
- We've kept the costs relatively low in productionalizing our models by running all automated jobs through RStudio using the Linux server CRONTAB.
RStudio is preferable to SPSS and SAS mainly because it is much lower cost to use even having a server license that we pay for the SAS licenses we used to have are too expensive and ultimately we decided to move away from using SAS for our reporting and data modeling needs.
RStudio doesn't really have support. Users search the web using StackOverflow for help in troubleshooting errors or for how to perform certain functions. RStudio also doesn't offer much in the way of training on the product, users are expecting to learn it more or less on the job or by taking separate classes that also use this product.
Do you think Posit delivers good value for the price?
Are you happy with Posit's feature set?
Did Posit live up to sales and marketing promises?
I wasn't involved with the selection/purchase process
Did implementation of Posit go as expected?
Would you buy Posit again?
Most data science applications are better handled in R and RStudio I find. I also appreciated that it is open-source software and is generally free to download and install new packages to stay up to date. It is less appropriate for data wrangling and manipulation for which I find there is no substitute for knowing SQL to extract and manipulate the data.