chptr1_third_update

I didn’t do that much coding since this update so, just going to use this one for our meeting.

• Notes for this update
  • 1. Had some difficulties coding these last two weeks. Didn’t quite get as far along as I had hoped, but I at least got some stuff mapped out. I wrote some more (I’m up to 4-5 pages now, and its helped me with my coding, I think). I created a new variable that examines how many times a census tract has had an increasing number of issued building perits from one 5 yr period to another. so the range is 0-3. This new variable brought the number of dimensions in the PCA from three to four, which I think is helpful for the LPA.
  • 2. I’ve been having a hard time getting separation of classes in the LPA. I think I may need to introduce another new variable to help with this. Hopefully we can talk about that at our meeting.
       
  • 3. Do you know anything about POMS (proportion of maximum)? I’ve found that when I use this for the LPA instead of scaling the variables, I get better separation of the classes. I’ve done some searching but I can’t figure out why its helping me out. Below are some results from both scaling and poms with the same data.
       
  • 4. I mapped out the results with the best separation in the three cities I’m examining. The results don’t make sense to me, which is why I think I may need to introduce another variable; I have some thoughts about that.
       
  • 5. see you in the morning.

# comparing models
#pca_orig_diminc <- as.data.frame(pca_orig_diminc)
dat_dim%>%
    dplyr::select(dim1,dim2,dim3,dim4) %>%
    single_imputation() %>%
    estimate_profiles(1:4, 
                      variances = c("equal", "varying"),
                      covariances = c("zero", "varying")
                      ) %>%
    compare_solutions(statistics = c("AIC", "BIC"))

## Compare tidyLPA solutions:
## 
##  Model Classes AIC      BIC      Warnings
##  1     1       19404.74 19445.11         
##  1     2       19415.26 19480.87         
##  1     3       18785.12 18875.96         
##  1     4       17374.01 17490.09 Warning 
##  6     1       19416.74 19487.39         
##  6     2       16420.67 16567.03         
##  6     3       15474.08 15696.14         
##  6     4       14910.16 15207.92         
## 
## Best model according to AIC is Model 6 with 4 classes.
## Best model according to BIC is Model 6 with 4 classes.
## 
## An analytic hierarchy process, based on the fit indices AIC, AWE, BIC, CLC, and KIC (Akogul & Erisoglu, 2017), suggests the best solution is Model 6 with 4 classes.

Scale vs POMS

Scale

y1 <- dat_dim %>% 
  dplyr::select(dim1,dim2,dim3,dim4) %>% 
  single_imputation() %>% 
  estimate_profiles(4,
                    variances = "varying",
                    covariances = "varying") %>% 
  plot_profiles()

y2<- dat_dim %>% 
  dplyr::select(dim1,dim2,dim3,dim4) %>% 
  single_imputation() %>% 
  estimate_profiles(4,package = "MplusAutomation",
                    variances = "varying",
                    covariances = "varying") %>% 
  plot_profiles()

POMS

y3 <- dat_dim %>% 
  dplyr::select(dim1,dim2,dim3,dim4) %>% 
  single_imputation() %>% 
  poms() %>% 
  estimate_profiles(4,
                    variances = "varying",
                    covariances = "varying") %>% 
  plot_profiles()

y4<- dat_dim %>% 
  dplyr::select(dim1,dim2,dim3,dim4) %>% 
  single_imputation() %>% 
  poms() %>% 
  estimate_profiles(4,package = "MplusAutomation",
                    variances = "varying",
                    covariances = "varying") %>% 
  plot_profiles()

Working with model y3

v3 <- dat_dim %>% 
  dplyr::select(dim1,dim2,dim3,dim4) %>% 
  single_imputation() %>% 
  poms() %>% 
  estimate_profiles(4,
                    variances = "varying",
                    covariances = "varying") %>% 
  get_data()


dat_dim$class <- v3$Class

table with numbers within the classes

table(v3$Class)

## 
##   1   2   3   4 
## 539 334 239  37

Mapping out the results in the three cities

Austin

Dallas

San Antonio

trying out five dimensions

pca_orig_diminc <- as.data.frame(pca_orig_inc$ind$coord[,1:5])
dat_dim5 <- cbind(bpsw_comb_inc,pca_orig_diminc)

dat_dim5 <- dat_dim5 %>% 
  mutate(
    dim1 = Dim.1,
    dim2 = Dim.2,
    dim3 = Dim.3,
    dim4 = Dim.4,
    dim5 = Dim.5
  )

# comparing models
#pca_orig_diminc <- as.data.frame(pca_orig_diminc)
dat_dim5%>%
    dplyr::select(dim1,dim2,dim3,dim4,dim5) %>%
    single_imputation() %>%
    estimate_profiles(1:5, 
                      variances = c("equal", "varying"),
                      covariances = c("zero", "varying")
                      ) %>%
    compare_solutions(statistics = c("AIC", "BIC"))

## Warning: 
## One or more analyses resulted in warnings! Examine these analyses carefully: model_1_class_5

## Warning: 
## One or more analyses resulted in warnings! Examine these analyses carefully: model_1_class_5

## Compare tidyLPA solutions:
## 
##  Model Classes AIC      BIC      Warnings
##  1     1       22779.16 22829.63         
##  1     2       22316.16 22396.90         
##  1     3       22146.09 22257.12         
##  1     4       21554.44 21695.75         
##  1     5       19867.71 20039.30 Warning 
##  6     1       22799.16 22900.10         
##  6     2       19397.94 19604.85         
##  6     3       18655.28 18968.17         
##  6     4       17571.80 17990.67         
##  6     5       17184.59 17709.44         
## 
## Best model according to AIC is Model 6 with 5 classes.
## Best model according to BIC is Model 6 with 5 classes.
## 
## An analytic hierarchy process, based on the fit indices AIC, AWE, BIC, CLC, and KIC (Akogul & Erisoglu, 2017), suggests the best solution is Model 6 with 5 classes.

y5 <- dat_dim5 %>% 
  dplyr::select(dim1,dim2,dim3,dim4, dim5) %>% 
  single_imputation() %>% 
  scale() %>% 
  estimate_profiles(5,
                    variances = "varying",
                    covariances = "varying") %>% 
  plot_profiles()

y6<- dat_dim5 %>% 
  dplyr::select(dim1,dim2,dim3,dim4,dim5) %>% 
  single_imputation() %>% 
  scale() %>% 
  estimate_profiles(5,package = "MplusAutomation",
                    variances = "varying",
                    covariances = "varying") %>% 
  plot_profiles()

## Warning: 
## One or more analyses resulted in warnings! Examine these analyses carefully: model_6_class_5

y7 <- dat_dim5 %>% 
  dplyr::select(dim1,dim2,dim3,dim4,dim5) %>% 
  single_imputation() %>% 
  poms() %>% 
  estimate_profiles(5,
                    variances = "varying",
                    covariances = "varying") %>% 
  plot_profiles()

y8<- dat_dim5 %>% 
  dplyr::select(dim1,dim2,dim3,dim4,dim5) %>% 
  single_imputation() %>% 
  poms() %>% 
  estimate_profiles(5,package = "MplusAutomation",
                    variances = "varying",
                    covariances = "varying") %>% 
  plot_profiles()

## Warning: 
## One or more analyses resulted in warnings! Examine these analyses carefully: model_6_class_5

trying to look at 5 dimensions with three classes

y9 <- dat_dim5 %>% 
  dplyr::select(dim1,dim2,dim3,dim4, dim5) %>% 
  single_imputation() %>% 
  scale() %>% 
  estimate_profiles(3,
                    variances = "varying",
                    covariances = "varying") %>% 
  plot_profiles()

y10<- dat_dim5 %>% 
  dplyr::select(dim1,dim2,dim3,dim4,dim5) %>% 
  single_imputation() %>% 
  scale() %>% 
  estimate_profiles(3,package = "MplusAutomation",
                    variances = "varying",
                    covariances = "varying") %>% 
  plot_profiles()

## Warning: 
## One or more analyses resulted in warnings! Examine these analyses carefully: model_6_class_3

y11 <- dat_dim5 %>% 
  dplyr::select(dim1,dim2,dim3,dim4,dim5) %>% 
  single_imputation() %>% 
  poms() %>% 
  estimate_profiles(3,
                    variances = "varying",
                    covariances = "varying") %>% 
  plot_profiles()

y12<- dat_dim5 %>% 
  dplyr::select(dim1,dim2,dim3,dim4,dim5) %>% 
  single_imputation() %>% 
  poms() %>% 
  estimate_profiles(3,package = "MplusAutomation",
                    variances = "varying",
                    covariances = "varying") %>% 
  plot_profiles()

## Warning: 
## One or more analyses resulted in warnings! Examine these analyses carefully: model_6_class_3

5 dimensions 4 classes

y11 <- dat_dim5 %>% 
  dplyr::select(dim1,dim2,dim3,dim4,dim5) %>% 
  single_imputation() %>% 
  poms() %>% 
  estimate_profiles(4,
                    variances = "varying",
                    covariances = "varying") %>% 
  plot_profiles()

y12<- dat_dim5 %>% 
  dplyr::select(dim1,dim2,dim3,dim4,dim5) %>% 
  single_imputation() %>% 
  poms() %>% 
  estimate_profiles(4,package = "MplusAutomation",
                    variances = "varying",
                    covariances = "varying") %>% 
  plot_profiles()

## Warning: 
## One or more analyses resulted in warnings! Examine these analyses carefully: model_6_class_4

5 dimensions 5 classes

y13 <- dat_dim5 %>% 
  dplyr::select(dim1,dim2,dim3,dim4,dim5) %>% 
  single_imputation() %>% 
  poms() %>% 
  estimate_profiles(5,
                    variances = "varying",
                    covariances = "varying") %>% 
  plot_profiles()

y14<- dat_dim5 %>% 
  dplyr::select(dim1,dim2,dim3,dim4,dim5) %>% 
  single_imputation() %>% 
  poms() %>% 
  estimate_profiles(5,package = "MplusAutomation",
                    variances = "varying",
                    covariances = "varying") %>% 
  plot_profiles()

## Warning: 
## One or more analyses resulted in warnings! Examine these analyses carefully: model_6_class_5