Regression Analysis

[Matthew Sutera]

Does anyone have any products that do a decent regression analysis directly from an MLS export? There are a ton available but all seem confusing to learn. Just want to make sure I choose a decent one if im going to put the work in learning it.

 

[Virginiageneral]

MS Excel! Just download your MLS info into an excel format. Make sure you have a large sample size. Then once you run the excel you can add or subtract various variables. For example, you might find that a 6th bathroom does not add any more credibility and may make the model ineffective. I'd try and keep your dataset as similar as possible. These models are good for some neighborhoods and homes. You'll need a lot of data and some good controls on your model, or else some of those emotional events (we have to buy this house, it's where I was born, or something stupid) can alter your results. Mass modeling has pros and cons, though. In a mass model, almost every comp just exported from MLS will never be as good as a comp that you have confirmed. For example, MLS may not report additional costs planned by the buyer, abnormal financing, the appliance package, etc.... If you have an admin, you can make a spreadsheet for each neighborhood you cover and everytime you add a good comp, have them add the property's variables into a spreadsheet. After a while, you'll have years of excellent data that you either confirmed by interviewing a buyer/seller/broker and not just a website printout that someone had their remote assistant from India update. We are only as good as our data!

 

[DWiley]

Does anyone have any products that do a decent regression analysis directly from an MLS export? There are a ton available but all seem confusing to learn. Just want to make sure I choose a decent one if im going to put the work in learning it.

This is the best one I have seen:

http://www.avttools.com/products/regression/

 

[Gobears81]

MS Excel! Just download your MLS info into an excel format. Make sure you have a large sample size. Then once you run the excel you can add or subtract various variables. For example, you might find that a 6th bathroom does not add any more credibility and may make the model ineffective. I'd try and keep your dataset as similar as possible. These models are good for some neighborhoods and homes. You'll need a lot of data and some good controls on your model, or else some of those emotional events (we have to buy this house, it's where I was born, or something stupid) can alter your results. Mass modeling has pros and cons, though. In a mass model, almost every comp just exported from MLS will never be as good as a comp that you have confirmed. For example, MLS may not report additional costs planned by the buyer, abnormal financing, the appliance package, etc.... If you have an admin, you can make a spreadsheet for each neighborhood you cover and everytime you add a good comp, have them add the property's variables into a spreadsheet. After a while, you'll have years of excellent data that you either confirmed by interviewing a buyer/seller/broker and not just a website printout that someone had their remote assistant from India update. We are only as good as our data!

Most of my statistical analysis is on my HP12C and that is obviously overly cumbersome. If I make a graph, I can't do it from scratch and have to input data on the template that was used to show my area's unemployment data every time. I'm only 35, but have the computer literacy of those that are 20-years older. Took my at least an hour to figure out how to put in a trendline recently. Tried to do a regression analysis in Excel on that graph that I put a trendline in and couldn't figure out how the output could be used to ascertain an implied value for my property.

Actually somewhat embarrassed to ask, but what is a good resource for making graphs and doing statistical analysis on Excel?

 

[Deleted member 130081]

Argh. Why do you want to employ a linear regression analysis? Let me guess, to support adjustments right? I'll say this, USPAP states we must reconcile the methods and techniques used in the appraisal. That means you have to stop and ask yourself if what you did was worth a turd or not. Have you ever considered that question with regression analysis as applied for adjustments? Garbage in, garbage out. If you don't have the appropriate source data, how can you expect a credible result? If you had the appropriate source data, you wouldn't need a regression analysis, unless you wanted to be fancy.

 

[jay trotta]

IMO- if in an area that offers sufficient data, it may be applicable, I do not have that luxury in most instances; I believe the question would be; when you drill down (closest city's near me would be a Block to Block market analysis) is there a sufficient amount of data in that tight area to extract information ??
Have run thru a few programs over the years and most required expansion out of a range I would consider reliable; think I just need to stick with the Old Mousetrap scenario for now. The requirements for; UAD / Lender / UW and reviewer, over the past ten years would lead me to believe it would be more of a time consuming effort and insufficient return for the time required.

 

[timd354]

Argh. Why do you want to employ a linear regression analysis? Let me guess, to support adjustments right? I'll say this, USPAP states we must reconcile the methods and techniques used in the appraisal. That means you have to stop and ask yourself if what you did was worth a turd or not. Have you ever considered that question with regression analysis as applied for adjustments? Garbage in, garbage out. If you don't have the appropriate source data, how can you expect a credible result? If you had the appropriate source data, you wouldn't need a regression analysis, unless you wanted to be fancy.

You need to open your horizons a little bit.

So how do you derive and support your adjustments? Let me guess, if you are like most other appraisers, you probably claim to utilize paired sale analysis and probably don't even realize that paired sales analysis is really a very simplistic and crude form of regression analysis that has a whole lot of limitations, not the least of which is trying to extrapolate a trend from an extremely limited data set (i.e., when you have only one or two paired sales, how do you know that these sales are not actually outliers?). Sure, regression analysis has its limitations and one has to be aware of those limitations, but regression analysis when properly done by a skilled appraiser is a very useful tool and can provide excellent support for adjustments in many circumstances.

 

[DWiley]

A common issue that I see is that regression seminars, even those aimed at appraisers, tend to teach the match of regression but not the practical application. I had an appraiser who used a fancy regression module for an ocean-front property in Florida, and, even though the cost approach indicated that the site accounted for over 50% of the value, the regression model did not include the site value as a variable. Hence, the results were pure garbage, at least with regard to supporting any line item adjustments.

 

[timd354]

A common issue that I see is that regression seminars, even those aimed at appraisers, tend to teach the match of regression but not the practical application. I had an appraiser who used a fancy regression module for an ocean-front property in Florida, and, even though the cost approach indicated that the site accounted for over 50% of the value, the regression model did not include the site value as a variable. Hence, the results were pure garbage, at least with regard to supporting any line item adjustments.

Like most tools in the appraiser's tool box, regression analysis is only as good as the user. However, far too many residential appraisers reflexively dismiss regression analysis because they don't understand regression, they don't have even a basic grasp of statistical analysis, they have never used regression, and it does take additional work and skill to be utilized properly.

 

[Deleted member 130081]

You need to open your horizons a little bit.

So how do you derive and support your adjustments? Let me guess, if you are like most other appraisers, you probably claim to utilize paired sale analysis and probably don't even realize that paired sales analysis is really a very simplistic and crude form of regression analysis that has a whole lot of limitations, not the least of which is trying to extrapolate a trend from an extremely limited data set (i.e., when you have only one or two paired sales, how do you know that these sales are not actually outliers?). Sure, regression analysis has its limitations and one has to be aware of those limitations, but regression analysis when properly done by a skilled appraiser is a very useful tool and can provide excellent support for adjustments in many circumstances.

I have opened my horizons, maybe too far even. It would be great to go back to the bliss of ignorance.

No, I do not claim to utilize matched pairs, as they are almost never available and for the same reason, regression analysis of adjustments is also not possible, at least not in any credible manner as far as I am concerned. Like you said, regression and paired sales for adjustments are essentially the same thing, where regression is just on a larger scale. If the data to do matched pairs isn't out there, how do you populate your regression model? Garbage in garbage out. A car is a great invention, but it will only run properly on gas, it wont run on plastic. Regression is no different.