Savvi’s data preparation technology
Written by a la mode on April 20, 2016
This is a guest post from the Savvi Analytics team. This is the fourth installment in a series from Savvi on some recent great improvements. Previous posts covered how Savvi works, their eBook on regression analysis, and the importance of balancing data. This week, the focus is on data preparation. You can find the Savvi automated regression analysis tool in the TOTAL Store here.
- What was the data screening process used to ensure that the data in which the regression model was developed was accurate and complete? Were properties with missing or invalid data identified and removed from the analysis and, if so, how many were removed?
- Prior to analyzing the property data, was the distribution of last known sales price checked for normality and were statistical outliers included or excluded from the analysis and on what basis (i.e., +- 2.0 standard deviations)?
- Were statistics such as MAPE provided with the appraisal regression analysis as evidence of the adequacy and accuracy of the overall regression model? Do the provided statistics indicate the model provides a good fit to the data?
- How many properties were examined in the regression analysis? Was there a minimum of at least 20 cases per each property characteristic included in the regression model to avoid shrinkage and missing one or more property characteristics potentially important to property values in a given market? Was bivariate regression or some other screening technique used to identify the property characteristics most likely associated with sales price before running the multivariate regression model?
As the old saying goes, "Garbage in, garbage out". Savvi makes managing your data a quick and easy process.