Apache Spark vs. SAS Enterprise Miner

Well suited: To most of the local run of datasets and non-prod systems - scalability is not a problem at all. Including data from multiple types of data sources is an added advantage. MLlib is a decently nice built-in library that can be used for most of the ML tasks. Less appropriate: We had to work on a RecSys where the music dataset that we used was around 300+Gb in size. We faced memory-based issues. Few times we also got memory errors. Also the MLlib library does not have support for advanced analytics and deep-learning frameworks support. Understanding the internals of the working of Apache Spark for beginners is highly not possible.

Incentivized

SAS Enterprise Miner is world-class software for individuals interested in developing reproducible models in a reasonable amount of time. Perhaps the most useful part of SAS Enterprise Miner is the ability to compare models with other models without writing code. The ensemble modeling capabilities is the easiest way to do ensemble modeling I have come across. SAS Enterprise Miner is well-suited for beginning to advanced analysts who know something about advanced analytics. The software is not well-suited for analysts or companies that have little interest in advanced modeling.

Incentivized

• Rich APIs for data transformation making for very each to transform and prepare data in a distributed environment without worrying about memory issues
• Faster in execution times compare to Hadoop and PIG Latin
• Easy SQL interface to the same data set for people who are comfortable to explore data in a declarative manner
• Interoperability between SQL and Scala / Python style of munging data

Incentivized

• Enterprise Miner is really visual and lets you do a whole lot without actually going into the detailed options. For decent results, you should really explore the different advanced options though.
• The recent versions of Miner allow users to use R code in Miner. You can then compare several models and approach to get the best performing model.
• The resulting data is really well displayed and easy to understand (ex: the lift graph, score ranking, etc.)
• Miner has the ability to integrate custom SAS code which allows the user to add functionalities that are specific to the project.

Incentivized

• Memory management. Very weak on that.
• PySpark not as robust as scala with spark.
• spark master HA is needed. Not as HA as it should be.
• Locality should not be a necessity, but does help improvement. But would prefer no locality

Incentivized

• SAS is not as user friendly as other stats software.

Incentivized

Capacity of computing data in cluster and fast speed.

If the team looking to use Apache Spark is not used to debug and tweak settings for jobs to ensure maximum optimizations, it can be frustrating. However, the documentation and the support of the community on the internet can help resolve most issues. Moreover, it is highly configurable and it integrates with different tools (eg: it can be used by dbt core), which increase the scenarios where it can be used

Incentivized

1. It integrates very well with scala or python. 2. It's very easy to understand SQL interoperability. 3. Apache is way faster than the other competitive technologies. 4. The support from the Apache community is very huge for Spark. 5. Execution times are faster as compared to others. 6. There are a large number of forums available for Apache Spark. 7. The code availability for Apache Spark is simpler and easy to gain access to. 8. Many organizations use Apache Spark, so many solutions are available for existing applications.

SAS' customer support used to be non-existent many years ago. Today, contacting SAS customer support is great. They are responsible, knowledgable, and seem to have an interest in getting the results right the first time. With that said, Enterprise Miner's online support is weak, probably because the user base is much smaller than other tools.

Incentivized

Spark in comparison to similar technologies ends up being a one stop shop. You can achieve so much with this one framework instead of having to stitch and weave multiple technologies from the Hadoop stack, all while getting incredibility performance, minimal boilerplate, and getting the ability to write your application in the language of your choosing.

Incentivized

SAS EM has a very great set of machine learning and predictive analytics toolsets, which helped our organization achieve its goals. We used other tools, but for us, SAS EM was the most intuitive and easy to learn the tool and it provides greater data exploration and data preparation capabilities compared to the other tools we used.

Incentivized

• Business leaders are able to take data driven decisions
• Business users are able access to data in near real time now . Before using spark, they had to wait for at least 24 hours for data to be available
• Business is able come up with new product ideas

Incentivized

• In our organization, users were using SAS already so the learning curve was really low. Within a few weeks after the implementation, the users were already delivering models developed with SAS Enterprise Miner. It is difficult to talk about ROI as models were already being developed before. It was mostly a change of technology and it was a smooth transition.
• Going with Enterprise Miner came with migration from desktop use of SAS to a server use of SAS. This created a new role of SAS administrator. This was obviously a cost but as the use of SAS increased greatly, it was expected.
• From a methodology standpoint, Enterprise Miner helped greatly in the documentation of the model development which was a requirement in a few groups such as the risk groups. Having a visual "GUI-like" approach to development, the flowchart or diagram of the project in Miner was able to give users a good understanding of the approach the analyst took to develop the model.

Incentivized