A great tool for predictive analytics and model development
March 30, 2019
A great tool for predictive analytics and model development
Score 8 out of 10
Vetted Review
Verified User
Overall Satisfaction with Alteryx
Within the Credit Risk Model development space, we use Alteryx to create time series models for PD, LGD and stress testing to compute VaR. It is also used by our predictive analytics department, but some use is spreading out to our data teams given the user-friendly interface and its ability to handle big chunks of GLM simulation for errors in data. It is also suited for a wide range of ETL things, from data cleansing to general data manipulation that helps answer a client's day-to-day questions. All in all, it's a great tool for working with advanced statistics.
- Seamless Data Input (Ingestion): Alteryx is capable of connecting with many data sources, we have used Tableau in cloud-hosted solutions with additional support of OleDB & ODBC connections.
- Advanced analytics: We have used statistical analytics tools which makes it easy for a user to look at data and provide simple summary statistics, while also having the ability to build advanced regression models.
- Macros: Just like other tools (e.g. SAS), Alteryx builds a macro. Be it a batch or iterative, it can save loads of efforts for standard and recursive processes. We've built numerous macros and use at least 10+ in every workflow post design so data is handled the same way in every workflow for consistency and cutting down on copy/pastes.
- Visualization needs severe improvement if it wants to be up to par from its market competitors like Tableau.
- The workflow scheduler has been reported as frozen in the earlier version which needs betterment.
- Lack of documentation for the few functionalities and higher cost is a trade-off as well.
To use advance statistics functions, we picked up alteryx with it's ease of use for model developers and users for testing who have lesser technology background with scripts directly imported from R or python. So it is simple to use and backward compatible with variety of data based ETL features as well.