Posit, formerly RStudio, is a modular data science platform, combining open source and commercial products.
N/A
Spotfire Data Science
Score 8.7 out of 10
N/A
Spotfire Data Science (formerly TIBCO Data Science) is a comprehensive platform for operationalizing data science, allowing users to scale data science across an organization to solve complex challenges faster and speed innovation. It is designed to enable data scientists to create innovative solutions using the latest machine learning techniques and open source developments. Create ML pipelines using a point-and-click UI or code. Orchestrate analytics using the tools, languages, and any…
In my humble opinion, if you are working on something related to Statistics, RStudio is your go-to tool. But if you are looking for something in Machine Learning, look out for Python. The beauty is that there are packages now by which you can write Python/SQL in R. Cross-platform functionality like such makes RStudio way ahead of its competition. A couple of chinks in RStudio armor are very small and can be considered as nagging just for the sake of argument. Other than completely based on programming language, I couldn't find significant drawbacks to using RStudio. It is one of the best free software available in the market at present.
It's well suited for companies who are working long-distance and who have lots of data to organize and present. My company does extensive field work collecting thousands of data points at a time, in many cases, so this is an excellent way to share data collected in the field with those still in the office so that reports can begin before our travels are really finished. It makes the office much more productive! Spotfire Data Science is less suited for companies with older employees or companies that are generally in one office and not doing distance work. It can be a little tricky for older people to grasp the interface and really utilize the program to its full potential so of course spending more time learning the software than using it effectively isn't a great use of time or money.
The support is incredibly professional and helpful, and they often go out of their way to help me when something doesn't work.
The one-click publishing from RStudio Connect is absolutely amazing, and I really like the way that it deploys your exact package versions, because otherwise, you can get in a terrible mess.
Python doesn't feel quite as native as R at the moment but I have definitely deployed stuff in R and Python that works beautifully which is really nice indeed.
The interface is screen based and much easier for teams of non-experts to use. Can, however, directly utilize in scripts and/or programming languages, so best of both worlds.
Has an extensive range of statistical methods, as well as extensive statistics around those methods.
Python integration is newer and still can be rough, especially with when using virtual environments.
RStudio Connect pricing feels very department focused, not quite an enterprise perspective.
Some of the RStudio packages don't follow conventional development guidelines (API breaking changes with minor version numbers) which can make supporting larger projects over longer timeframes difficult.
Although the graphics can be copied into Word or Excel, the copying and pasting process is buggy. Sometimes I basically have to do a Capture Rectangle of a graphic and paste it as a jpg, rather than having an embedded and editable graphic in my Word document.
The dialog boxes for some of the common tests, like multiple regression and ANOVA, use quirky terminology that is not intuitive to new users coming out of grad school with a more updated education. I have been using it for over ten years, and the dialogs have not updated in all of that time.
The tests like NMDS and PCA are clunky and difficult to use and get the appropriate outputs, especially compared to specialized software like PRIMER.
There is no viable alternative right now. The toolset is good and the functionality is increasing with every release. It is backed by regular releases and ongoing development by the RStudio team. There is good engagement with RStudio directly when support is required. Also there's a strong and growing community of developers who provide additional support and sample code.
The company is hesitant to spend this much on software. They are primarily an engineering firm, and they don't understand the use of analytical software for environmental professionals.
I think it's a quick and easy to use tool. The IDE is very intuitive and easy to adapt to. You do not need to learn a lot of things to use this tool. Any programmer and a person with knowledge or R can quick use this tool without issues.
RStudio is very available and cheap to use. It needs to be updated every once in a while, but the updates tend to be quick and they do not hinder my ability to make progress. I have not experienced any RStudio outages, and I have used the application quite a bit for a variety of statistical analyses
Since R is trendy among statisticians, you can find lots of help from the data science/ stats communities. If you need help with anything related to RStudio or R, google it or search on StackOverflow, you might easily find the solution that you are looking for.
RStudio was provided as the most customizable. It was also strictly the most feature-rich as far as enabling our organization to script, run, and make use of R open-source packages in our data analysis workstreams. It also provided some support for python, which was useful when we had R heavy code with some python threaded in. Overall we picked Rstudio for the features it provided for our data analysis needs and the ability to interface with our existing resources.
I prefer Spotfire Data Science's approach. It is more natural and fits the way I think. I prefer to use Spotfire Data Science's VB for writing macros. It is real code, meaning that I do not need to trick the software to do what I need and there are no implied loops over solving simple problems. The graphs are publication quality and can be edited by hand or using a macro if I am building hundreds of them. Spotfire Data Science had a user-friendly approach to building lengthy data processing streams (in its workspaces). It is just so fast for analyzing a dataset that you have never seen before and efficient for ongoing work on the same data.
RStudio is very scalable as a product. The issue I have is that it doesn't necessarily fit in nicely with the mainly Microsoft environment that everybody else is using. Having RStudio for us means dedicated servers and recruiting staff who know how to manage the environment. This isn't a fault of the product at all, it's just part of the data science landscape that we all have to put up with. Having said that RStudio is absolutely great for running on low spec servers and there are loads of options to handle concurrency, memory use, etc.
Using it for data science in a very big and old company, the most positive impact, from my point of view, has been the ability of spreading data culture across the group. Shortening the path from data to value.
Still it's hard to quantify economic benefits, we are struggling and it's a great point of attention, since splitting out the contribution of the single aspects of a project (and getting the RStudio pie) is complicated.
What is sure is that, in the long run, RStudio is boosting productivity and making the process in which is embedded more efficient (cost reduction).
Our company has had the program for less than 1 year. We don't expected a positive return this year. The goal is for Data Science to led to defined projects by the end of the end of the year and implementation in the following two. Overall, we are planning on 4 years to fully recoup the cost of the software and the cost of implementing identified projects.
