FHA Loan Interest Rate Model App
Vadim Bondarenko
May, 15, 2016
The App's Purpose
Allows users to explore the relationship between the interest rate on newly issued FHA loans and various loan, borrower, and demographic variables as predictors.
- Uses data from US Housing and Urban Development (HUD) website
- Fits multivariate linear model with Interest Rate as dependent variable
- Allows users to select a US state and various independent variables
- Reports model results, confidence intervals, and residual diagnostics
The Data
summary(dt_mod)
treas_30yr mortg_30yr zip_med_income zip_pct_white
Min. :2.460 Min. :3.350 Min. : 34060 Min. :0
1st Qu.:2.880 1st Qu.:3.570 1st Qu.: 35602 1st Qu.:0
Median :3.040 Median :3.890 Median : 59840 Median :0
Mean :3.096 Mean :3.867 Mean : 61672 Mean :0
3rd Qu.:3.260 3rd Qu.:4.070 3rd Qu.: 72312 3rd Qu.:0
Max. :3.890 Max. :4.490 Max. :154444 Max. :1
NA's :4561206 NA's :4561206
rate amount rate_type down_pmt_src
Min. :0.250 Min. : 2501 Length:4566831 Length:4566831
1st Qu.:3.625 1st Qu.: 115862 Class :character Class :character
Median :3.875 Median : 159442 Mode :character Mode :character
Mean :3.915 Mean : 182316
3rd Qu.:4.250 3rd Qu.: 225077
Max. :8.250 Max. :1427959
purpose property year
Length:4566831 Length:4566831 2016: 89426
Class :character Class :character 2012:1281951
Mode :character Mode :character 2013:1200197
2014: 779453
2015:1215804
lender
Other :4233490
WELLS FARGO BANK NA: 333341
Example Plot
mod_lm <- lm(rate ~ amount+purpose+mortg_30yr+zip_med_income, data = dt_mod)
texreg::plotreg(mod_lm)
