project_graph_sketches.Rmd

---
title: "Project graph sketches"
author: "J. Hamski"
date: "May 8, 2016"
output: html_document
---

```{r}
library(ggplot2)
library(dplyr)
library(tidyr)
```

```{r}
load("ConsumerExpectationsExplorer/data_demo.Rda")
load("ConsumerExpectationsExplorer/data_all.Rda")
load("ConsumerExpectationsExplorer/survey_microdata.Rda")
```

```{r}
library(ggbiplot)
```


```{r}
question.PCA <- unique(data.demo$Question)[1]

PCA.dataset <- data.demo %>%
  unite(col = Survey, Question, Demographic) %>%
  spread(key = Survey, value = results) %>% 
  select(-date) %>%
  select(matches(question.PCA)) %>% 
  as.matrix() %>%
  prcomp(tol = sqrt(.Machine$double.eps), scale = T)
```

```{r}
ggbiplot(PCA.dataset, obs.scale = 1, var.scale = 1, circle = TRUE) 
```




```{r}
PCA.dataset2 <- data.all %>%
  unite(col = Survey, Question, survey) %>%
  spread(key = Survey, value = results) %>% 
  select(-date) %>%
  as.matrix() %>%
  prcomp()
```

```{r}
ggbiplot(PCA.dataset2, obs.scale = 1, var.scale = 1, circle = TRUE)

```


```{r}
require(caret)
trans = preProcess(PCA.dataset, 
                   method=c("BoxCox", "center", 
                            "scale", "pca"))
PC = predict(trans, PCA.dataset)
```

```{r}
ggbiplot(PC, obs.scale = 1, var.scale = 1, circle = TRUE)

```




```{r}
seasonal.decom.data <- data.all %>%
  filter(Question == "Inflation expectations") %>%
  filter(survey == "Median one-year ahead expected inflation rate") %>%
  select(results) 

seasonal.decom.data <- ts(seasonal.decom.data$results, deltat = 1/12)
```

```{r}
stl.series <- stl(seasonal.decom.data, s.window="periodic")
```




```{r}
library(caret)
```

```{r}
featurePlot(x = microdata[,c(78,85, 88)],
            y = microdata[,162], 
            plot = "pairs",)
```


```{r}
source('project_theme.R')
```



Factor

```{r}
column.1 <- "`Financially better or worse off 12 months from now`"

ggplot(microdata) + geom_bar(aes_string(x = column.1), stat = "count") + project.theme()
```


Number

```{r}
column.2 <- "`Rate of inflation / deflation over next 12 months`"

ggplot(microdata, aes_string(x = column.2)) + geom_density() + xlim(-100, 100)
```

```{r}
q = quantile(as.matrix(select_(microdata, column.2)), na.rm = TRUE, probs =  c(0.1, 0.9))
```

```{r}
 ifelse(typeof(select_(microdata, column.1)[[1]]) == "double", "yes", "no")
```


All Repondents vs Demographic Comparisons
```{r}
survey.input = "Median one-year ahead expected inflation rate"
demo.input.question = "Median One Year Ahead Inflation Rate Expectation"
demo.input.demo = "Education High School or Less"

#values <- c("All Respondents" = "blue", demo.input.demo[1] = "red")

ggplot() + 
  geom_line(data =  filter(data.all, survey == survey.input), aes(x = date, y = results, color = "All Respondents")) + 
  geom_line(data =  filter(data.demo, Question == demo.input.question & Demographic == demo.input.demo), aes(x = date, y = results, color = as.character(demo.input.demo))) + 
  scale_color_manual(values = c("blue", "red"))

```