03-data-transformation.Rmd

---
title: "Data transformation"
output: html_notebook
---
 
## Class catchup

```{r}
library(dplyr)
library(dbplyr)
library(DBI)
con <- dbConnect(odbc::odbc(), "Postgres Dev")
airports <- tbl(con, in_schema("datawarehouse", "airport"))
flights <- tbl(con, in_schema("datawarehouse", "vflight"))
carriers <- tbl(con, in_schema("datawarehouse", "carrier"))
```

##  3.1 - Group and sort records

1. How many flights are there per month?
```{r}
flights %>%
  group_by(month) %>%
  tally() 
```

2. Order the results by the month number by using `arrange()`
```{r}
flights %>%
  group_by(month) %>%
  tally() %>%
```

3. Order the results by the number of flights, starting with the month with most flights by using `desc()` inside the `arrange()` command
```{r}

```

## 3.2 - Answering questions with `dplyr`

1. Which are the top 4 months with the most flight activity?
```{r}
flights %>%
  group_by(month) %>%
  tally() %>%
  arrange(desc(n)) %>%
  head(4)
```

2. What were the top 5 calendar days with most flight activity?
```{r}

```

3. Which are the top 5 carriers (airlines) with the most flights?

```{r}

```

4. Figure the percent ratio of flights per month
```{r}
flights %>%
  group_by(month) %>%
  tally() %>%
  arrange(desc(n)) %>%
  mutate(percent = n/sum(n, na.rm = TRUE))
```

5. Figure the percent ratio of flights per carrier
```{r}

```

## 3.3 - Aggregate mulitple columns

1. Use `summarise_all()` to send the same function to all fields
```{r}
flights %>%
  select(depdelay, arrdelay) %>%
  summarise_all(mean, na.rm = TRUE)
```

2. Use `summarise_at()` to pre-select the fields that will receive the function
```{r}
flights %>%
  summarise_at(c("depdelay", "arrdelay"), mean, na.rm = TRUE)
```

3. Use `summarise_if()` to summarize only if the field meets a criterion
```{r}
flights %>%
  summarise_if(is.numeric,mean, na.rm = TRUE)
```

4. Combine with `group_by()` to create more complex results
```{r}
flights %>%
  select(month, depdelay, arrdelay) %>%
  group_by(month) %>%
  summarise_all(mean, na.rm = TRUE)
```

## 3.4 - View record level data

How many flights in July 18th were one or more hours late?
```{r}
flights %>%
  filter(
    depdelay >= 60,
    month == 7,
    dayofmonth == 18
  ) %>%
  tally()
```

1. Use `filter()` to retrieve only the needed data, and `head()` to limit the preview even further.
```{r}
flights %>%
  filter(
    depdelay >= 60,
    month == 7,
    dayofmonth == 18
  ) %>%
  head(100)
```

2. Use `collect()` and `View()` to preview the data in the IDE. Make sure to **always** limit the number of returned rows. https://github.com/tidyverse/tibble/issues/373

```{r, eval = FALSE}
flights %>%
  filter(
    depdelay >= 60,
    month == 7,
    dayofmonth == 18
  ) %>%
  collect() %>%
  head(100) %>%
  View("my_preview")
```

## 3.5 - Case statements

1. Use `case_when()` to bucket each month one of four seasons
```{r}
flights %>%
  mutate(
    season = case_when(
      month >= 3 && month <= 5  ~ "Spring",
      month >= 6 && month <= 8  ~ "Summer",
      month >= 9 && month <= 11 ~ "Fall",
      TRUE ~ "Winter"
    )
  ) %>%
  group_by(season) %>%
  tally()
```

2. Add a specific case for "Winter"
```{r}

```

3. Append an entry for Monday at the end of the case statement
```{r}

```

4. Move the "Monday" entry to the top of the case statement
```{r}

```


## 3.6 - Data enrichment

1. Load the `planes` data into memory
```{r}
planes <- nycflights13::planes
```

2. Using `DBI`, copy the `planes` data to the datawarehouse as a temporary table, and load it to a variable
```{r}
dbWriteTable(con, "planes", planes, temporary = TRUE)
tbl_planes <- tbl(con, "planes")
```

3. Create a "lazy" variable that joins the flights table to the new temp table
```{r}
combined <- flights %>%
  left_join(tbl_planes, by = "tailnum") 
```

4. View a sample of flights of planes with more than 100 seats
```{r}
combined %>%
  filter(seats > 100) %>%
  head()
```

5. How many flights are from McDonnel Douglas planes 
```{r}
combined %>%
  filter(manufacturer == "MCDONNELL DOUGLAS") %>%
  
```

6. See how many flights each plane (tailnum) McDonnel Douglas had
```{r}

``` 

7. Get the total number of planes, and the average, minimum & maximum number of flights for McDonnel Douglas
```{r}

``` 

8. Disconnect from the database
```{r}
dbDisconnect(con)
```