Hi,
I would like to construct brand preference map using output from Factor analysis? I have 5 brands and 11 factors. How can I get X and Y coordinates for each factor and brand ? I have never done that and would much appreciate if you can suggest some ways? I am using SPSS and Excel. I have attached an example of the map I would like to construct
Best regards,
Davit
A simple and solid way to start facing a situation like the one you are describing is to rely on a linear analysis (variations within a range, percentages, probability, etc.). Note that these ideas are applicable to any situation where, out of a list of common variables, you have to determine which set of values is better.
By assuming that you have brand A (factor1=2 and factor2=5) and brand B (factor1=3 and factor2=3), the way to proceed would be:
- Determine the min/max range for each factor. That is: factor1 -> (2, 3) and factor2 -> (3, 5).
- For each brand, calculate the relationship of each factor with respect to the corresponding range. A simple and effective approach would be: (value-min)/(max-min). brand A -> (2-2)/(3-2) and (5-3)/(5-3); brand B -> (3-2)/(3-2) and (3-3)/(5-3).
- To get a global assessment for each brand, you would have to aggregate all the values for all the factors. The most straightforward approach is to just add them; eventually, you might also include weights indicating differences in the importance of the factors; brand A -> weight1*0 + weight2*1; brand B -> weight1*1 + weight2*0.
I think that this is a very solid (and easily extensible) first step to analyse scenarios of this kind.
Bear in mind that this is just a generic proceeding to determine the best options out of a set of inputs; nothing to do with a graph, although creating one from this information would be very easy. Also note that there might be in-built functionalities outputting direct answers for your exact conditions right away, but I thought that these ideas would be more helpful.
Dear Alvaro,
Thank you for your comprehensive answer.
You mentioned that it would be easy to get a graph after I follow those steps. So can I get X and Y coordinates for each brand? if yes, can you a bit elaborate on that?
I know that for example XLSTAT gives exactly what I intent to do, however I am working with SPSS and excel so I need to make some manipulations manually.
Looking forward to hearing from you
Best regards,
Davit
Your initial data doesn't provide the answer you are after (i.e., which brand is better?) right away and, consequently, no descriptive-enough graph is immediately possible. Even in-built functionalities delivering meaningful graphs would have to perform internal calculations on the lines of the ones I suggested.
On the other hand, if you follow the steps I proposed (or any other version delivering a clear comparison), you would get values which readers could immediately understand, through graphs or in any other way. For example, by assuming brand A having 1.23 and brand B 1.5, virtually any 2D graph (e.g., X for brands and Y for values) could display the information in a meaningful way: readers would be able to quickly tell which brand is better and for how much.
There are many ways to perform the comparison and to display the results. As said, take my comments just as some preliminary ideas helping you to approach this situation comfortably.