Overall Satisfaction with RapidMiner Studio
We use RapidMiner to create ETL (Extract, Transform, Load) processes to load our BI datamarts with data from operational databases. We've created complex load processes, and we prepare that data to be fed, and later on, create Business Intelligence dashboards. We also use RapidMiner to perform some data mining with different techniques such as text processing, image processing, and algorithm data analysis (clustering, neuralnet, correlation, etc...)
- RapidMiner is really fast at reading all kinds of databases. We read and merge databases like SQL Server, Informix, MySQL, and Oracle. Configuring access is easy, some drivers are inbuilt, but it's not difficult to find new java drivers to allow RapidMiner to connect to other databases.
- Performing all kinds of transformations, calculations (date, percentages...), joins, and filters without coding. We have several different databases and this makes my life a lot easier. Knowing that this part is 80% of analyst work, you know that you can work more on the analyses itself and not on cleaning and preparing data.
- You can clone transformations to reuse on new analyses, so you save a lot of time. There's a lot of add-ons to make different things (text, image analysis, recommender systems, etc).
- Training is easy, the tool is intuitive and there's a lot of videos on the internet. The community is very active.
- Sharing RapidMiner Studio analysis is not easy. You may think that the RapidMiner Server does that work but no. It's more automated job oriented or useful to run models on a web site. If you need to use it for Business Analytics dashboards, this is not the tool. It's more a backstage tool for analyst. Some charts are good but other not so much.
- The free edition allows you to work with 10,000 rows, but if you need more, it's not cheap (100,000 rows - 2,500 USD/year, 1,000,000 rows - 5,000 USD/year).
- The commercial team is not very reactive. I've asked for a RapidMiner Education Program and Rapidminer Server quotation with no answer. I guess that's because they were changing from an opensource company model to a more commercial one.
- Very high positive impact because it's very fast to work with data with no coding need. You save a lot of time.
For me, the best advantage to use RapidMiner is the ease of use to learn and deploy new processes. Yo don't need to code, you learn fast and it's really flexible when it comes to transforming data. Knime is also good, but not so flexible, and visually less attractive. Pentaho ETL is less intuitive although very powerful but it will do only ETL part without mining techniques. SQL Server is way more difficult to use and deploy and not as powerful for ETL purposes. SQL Server has some good capabilities with using predictive algorithms.
RapidMiner Feature Ratings
Using RapidMiner Studio
2 - Data analysts
- ETL process and data analysis
- Databases access
- Mining data, algorithm analysis
- Natural language analytics, image mining
Using RapidMiner Studio
Pros | Cons |
---|---|
Like to use Relatively simple Technical support not required Well integrated Consistent Quick to learn Convenient Feel confident using Familiar | None |
- Apply multiple transformations to data (almost unlimited!)
- Ease to include new functionalities
- A lot of templates to learn and modify for your own needs
- Some algorithm analysis are not very intuitive (clusters for example)
- Charts needs to get better
- Sharing analysis is not well done