class_3.Rmd

---
title: "in class coding"
author: "Julia Gallucci"
date: "13/12/2023"
output: html_document
---
Load our libraries
```{r}
#install.packages("haven")
library(haven)
library(tidyverse)
```

Read in our datasets
```{r}
ads_data <- read_sav("./ads_raw.sav")
ces_2019_raw <- read_csv("./ces_2019_raw.csv")
```

Let's have a look!
```{r}
glimpse(ads_data)
```

Filter our data
```{r}
filter(ads_data, Duration__in_seconds_ < 100)
```
Arrange our data
```{r}
arrange(ads_data, Duration__in_seconds_) #ascending
arrange(ads_data, -Duration__in_seconds_) #descending
```
Select specific columns from our data
```{r}
select(ads_data, RecordedDate) #select columns
select(ads_data, -Consent, -DistributionChannel) #remove columns
```
Multiple operations using the pipe operator
```{r}
ads_data %>%
  filter(Duration__in_seconds_ < 100) %>%
  arrange(Duration__in_seconds_) %>%
  select(RecordedDate, Duration__in_seconds_)
```
Mutate columns in our data
```{r}
ads_data <- ads_data %>%
  mutate(Birthyear_add_day = 
           str_c(Birthyear, "07-01")) %>%
  mutate(Birthyear_add_day =
           as_datetime(Birthyear_add_day))
ads_data$Birthyear_add_day
```

```{r}
ads_data <- ads_data %>%
  mutate(age = EndDate - Birthyear_add_day)

ads_data %>%
  select(age)
```
Summarize our data
```{r}
summary(ads_data)
```

Pulling a variable from our dataset
```{r}
#MEDIAN
ads_data  %>%
  pull(Duration__in_seconds_) %>%
  median(na.rm = TRUE)
#MEAN
ads_data %>%
  pull(Duration__in_seconds_) %>%
  mean(na.rm = TRUE)
#RANGE
ads_data %>%
  pull(Duration__in_seconds_) %>%
  range(na.rm = TRUE)
#VARIANCE
ads_data %>%
  pull(Duration__in_seconds_) %>%
  var(na.rm = TRUE)
#STANDARD DEVIATION
ads_data %>%
  pull(Duration__in_seconds_) %>%
  sd(na.rm = TRUE)
```

Summarise our data
```{r}
ads_data %>%
  summarise(mean_time = mean(Duration__in_seconds_, na.rm = TRUE),
            sd_time = sd(Duration__in_seconds_, na.rm = TRUE))
```
Grouping based on categorical variable
```{r}
ads_data %>%
  group_by(Gender) %>%
  summarise(counts = n(),
            mean_time = mean(Duration__in_seconds_, na.rm = TRUE),
            sd_time = sd(Duration__in_seconds_, na.rm = TRUE))
```

Data cleaning
```{r}
CES_data <- ces_2019_raw %>%
  mutate(cps19_yob_fix = cps19_yob +1919)

CES_data %>%
  pull(cps19_yob_fix) %>%
  range(na.rm = TRUE)
```

```{r}
CES_data <- CES_data %>%
  mutate(age = 2019 - cps19_yob_fix)

CES_data %>%
  pull(age) %>%
  range(na.rm = TRUE)

CES_data <- CES_data %>%
  mutate(cps19_gender_fix = factor(cps19_gender)) %>%
  mutate(cps19_gender_fix =
           fct_recode(cps19_gender_fix,
                      "M" = "1",
                      "F"= "2",
                      "NB" = "3"
                      ))
CES_data %>%
  select(cps19_gender, cps19_gender_fix)
```

```{r}
CES_data %>%
  filter(cps19_household > 10) %>%
  arrange(-cps19_household) %>%
  pull(cps19_household)
```
```{r}
CES_data <- CES_data %>%
  mutate(cps19_household = ifelse(cps19_household > 15, NA, cps19_household))

CES_data %>%
  filter(cps19_household > 10) %>%
  pull(cps19_household)
```
```{r}
CES_data %>%
  filter(cps19_income_number > 1000000) %>%
  arrange(-cps19_income_number) %>%
  pull(cps19_income_number)
```
```{r}
CES_data <- CES_data %>%
  mutate(cps19_income_number = ifelse(cps19_income_number >= 100000000, NA, cps19_income_number)) 
CES_data %>%
  filter(cps19_income_number > 1000000) %>%
  pull(cps19_income_number)
```
Data summarizing
```{r}
CES_data <- ces_2019_cleaned
CES_data %>%
  filter(cps19_province == "Ontario") %>%
  select(cps19_prov_gov_sat,
         cps19_prov_id,
         cps19_income_number)
```
```{r}
CES_data %>%
  filter(cps19_province == "Ontario") %>%
  select(cps19_prov_gov_sat,
         cps19_prov_id,
         cps19_income_number) %>%
  group_by(cps19_prov_id) %>%
  summarise(median_income = 
              median(cps19_income_number, na.rm = TRUE), count = n())
```
```{r}
CES_data %>%
  filter(cps19_province =="Ontario") %>%
  select(cps19_prov_gov_sat,
         cps19_prov_id,
         cps19_income_number) %>%
  mutate(cps19_prov_id = factor(cps19_prov_id, levels = c("Liberal","Progressive Conservative","NDP","Green","Another party","None", "Don't know/prefer not to answer"))) %>%
  group_by(cps19_prov_id) %>%
  summarise(median_income = median(cps19_income_number, na.rm = TRUE), count = n())
```

```{r}
CES_data %>%
  filter(cps19_province == "Ontario") %>%
  select(cps19_prov_gov_sat,
         cps19_prov_id,
         cps19_income_number) %>%
  group_by(cps19_prov_gov_sat) %>%
  summarise(median_income = median(cps19_income_number, na.rm = TRUE), count = n())
```

```{r}
CES_data %>%
  filter(cps19_province == "Ontario") %>%
  select(cps19_prov_gov_sat,
         cps19_prov_id,
         cps19_income_number) %>%
  group_by(cps19_prov_id) %>%
  summarise(median_income = median(cps19_income_number, na.rm = TRUE), count = n()) %>%
  arrange(-count)
```
```{r}
CES_data %>%
  filter(cps19_province == "Ontario") %>%
  select(cps19_prov_gov_sat,
         cps19_prov_id,
         cps19_income_number) %>%
  mutate(cps19_prov_id = factor(cps19_prov_id, levels = c("Liberal","Progressive Conservative","NDP","Green","Another party","None","Don't know/prefer not to answer"))) %>%
  mutate(cps19_prov_gov_sat = factor(cps19_prov_gov_sat, levels = c("Not at all satisfied","Not very satisfied", "Fairly satisfied","Very satisfied","Don't know/prefer not to answer"))) %>%
  group_by(cps19_prov_gov_sat, cps19_prov_id) %>%
  summarise(median_income = median(cps19_income_number, na.rm = TRUE)) %>%
  spread(key = cps19_prov_gov_sat,
         value = median_income)
```













Exercises
1. Filter the rows in the CES_data where survey taker is between 30 and 50 years old 
```{r}
CES_data %>%
  filter(cps19_age > 30 & cps19_age < 50) %>%
  select(cps19_age) %>%
  pull(cps19_age) %>%
  range()
```

2. Filter the rows in the CES_data where the survey taker answered the cps19_votechoice question (not NA)
```{r}
CES_data %>%
  filter(!(is.na(cps19_votechoice))) %>%
  select(cps19_votechoice)
```

3. Selecting the columns cps19_age and cps19_province from CES_data
```{r}
CES_data %>%
  select(cps19_age, cps19_province)
```

4. Select all but cps19
```{r}
CES_data %>%
  select(-cps19_province)
```

1. Create new variable in CES_data that tells us whether or not a person consumes news content (cps19_news_cons == 0 minutes OR not)
```{r}
CES_data %>%
  mutate(news_watcher = ifelse(cps19_news_cons == "0 minutes", "Not watcher", "Watcher")) %>%
  select(cps19_news_cons, news_watcher)
```

2. Modify the variable cps19_income_number so that its measured in 1000s 
```{r}
CES_data %>%
  mutate(cps19_income_number = cps19_income_number / 1000)
```

1. CES_data data; group by cps19_votechoice. find median and mean rating for Trudeau
```{r}
CES_data %>%
  select(cps19_lead_rating_23, cps19_votechoice) %>%
  group_by(cps19_votechoice) %>%
  summarise(median_rating = median(cps19_lead_rating_23, na.rm = TRUE), mean_rating = mean(cps19_lead_rating_23, na.rm = TRUE))
```
2. Use the CES_data dataset. Group by cps19_imm and cps19_spend_educ. Find the count for each group.
```{r}
CES_data %>%
  select(cps19_imm, cps19_spend_educ) %>%
  group_by(cps19_imm, cps19_spend_educ) %>%
  summarise(counts = n()) %>%
  spread(key = cps19_imm, value = counts)
```