We often talk to people who are unsure whether they need SPSS Statistics or whether SPSS Modeler might be more suited to their needs. In fact, it’s not always a clear cut choice as to which tool is more appropriate as it depends on the context in which the technology might be used. With that in mind I thought it might be helpful to develop a little infographic to lay out the sorts of things that you should be thinking about when choosing between SPSS Modeler and SPSS Statistics.
We can think of the choice as a sort of continuum, where you may lean more towards one product or the other depending on how you answer the following questions.
- What are your models for? If you’re building models where the aim is to generate outcomes for operational decisions, then SPSS Modeler is very strong in this area. Whereas if you’re developing models purely for research and insight purposes you might find that the classical multivariate procedures in SPSS Statistics are more appropriate.
- What’s the focus of your modelling? If you’re really interested in accurate prediction then SPSS Modeler has excellent functionality for quickly building, comparing, testing and combining predictive models. If however your analysis is more about understanding variation from a statistical standpoint and particularly if you need to carry out significance tests, then SPSS Statistics might be the better choice.
- What was the original purpose of your data? If your data was specifically collected with analysis in mind then SPSS Statistics may be ideal. If however, you need to analyse data from customer databases and flat files that were originally collected with marketing, billing or CRM applications in mind, you might find that SPSS Modeler is more adept and consolidating and cleaning these data.
- How many sources of data are you working with? It’s easier to consolidate data from multiple data sources and in multiple file formats in SPSS Modeler than it is in SPSS Statistics. If you’ve just got flat files or data from a single source, then SPSS Statistics may be sufficient for your needs.
- Will you be analysing data on a fairly ad hoc basis? Modeler is more appropriate for analysts that want to develop re-usable analytical applications (like fraud detection) whereas SPSS Statistics is often used to analyse new data files as and when they appear.
- Do you need to create regular analytical reports? SPSS Statistics is ideal for creating analytically-driven reports as it has great tabulation and charting capabilities as well as the ability to save jobs as SPSS syntax so they can be applied to updated data. Modeler on the other hand, is more commonly used for ‘pattern detection’ type problems than traditional reporting.