My experience with SQLServer and its R Integration
April 02, 2018
My experience with SQLServer and its R Integration
Score 8 out of 10
Vetted Review
Verified User
Overall Satisfaction with Microsoft SQL Server
We used to SQL Server to build a decision support system for mid-size companies (e.g. bakeries) offering perishable goods. The core of the system is a machine learning algorithm that provides demand forecast based on transactional data (POS data) that are required for daily production and order decisions. Hence, the provided demand forecasts are used by operational staff at the headquarters that control the complete supply chain.
- Integration of the R programming language and its eco-system.
- Transact-SQL is quite expressive and allows implementing complex application logic.
- MS Excel can be used as front-end.
- OLAP queries on non-aggregated POS data can be very slow. Hence, it is required to persist aggregated views on the data.
- Debugging R-Code is not convenient.
- Versioning of SQL procedures.
- Decision support leads to better inventory management of perishable items.
- Provided demand forecasts lead to time saving of operational staff.
- Better product availability causes a revenue increase.
The pricing model of Microsoft and the integration of R makes it an appropriate choice for many applications that are not based on a very large data foundation. In particular, the integration of R allows the deployment of machine learning models within the database. Moreover, it is easily possible to connect Excel to the database as front-end for various reports.