lab3.qmd

---
title: "Lab 3"
author: "Carissa Feliciano"
format: html
embed-resources: true
---

# 1. Read in the data
```{r}
download.file(
  "https://raw.githubusercontent.com/USCbiostats/data-science-data/master/02_met/met_all.gz",
  destfile = file.path("~", "Downloads", "met_all.gz"),
  method   = "libcurl",
  timeout  = 60
)

met <- data.table::fread(file.path("~", "Downloads", "met_all.gz"))
met <- as.data.frame(met)
```
# 2. Check the dimensions, headers, and footers.

```{r}
dim(met)
```
There are 2377343 rows and 30 columns.

```{r}
head(met)
```

```{r}
tail(met)
```
# 3. Take a look at the variables.

```{r}
str(met)
```

The key variables related to our question of interest are elev, wind.sp, temp, year, day, and hour. 

# 4. Take a closer look at the key variables. 
```{r}
table(met$year)
```

```{r}
table(met$day)
```

```{r}
table(met$hour)
```

```{r}
summary(met$temp)
```
```{r}
summary(met$elev)
```

```{r}
summary(met$wind.sp)
```
```{r}
met[met$elev==9999.0, "elev"] <- NA
summary(met$elev)
```
The highest weather station is at 4113 meters about mean sea level.

```{r}
met <- met[met$temp > -40, ]
head(met[order(met$temp), ])
```

```{r}
summary(met$wind.sp)
```

For the wind.sp variable, there are 91832 missing values. 

# 5. Check the data against an external source.
```{r}
met_ss <- met[met$temp == -17.2, c('hour','lat','lon','elev','wind.sp')]

summary(met_ss)
```
The suspicious temperature value (-17.2C) is located at (38.77,-104.3), which is outside of Colorado Springs, Colorado. The closest town is Truckton, Colorado, which is at the same elevation. The average climate in August is a high of 82F and low of 57F (13.88C). The value of -17.2C does not seem reasonable in this context. 

![38.77, -104.3](-17.2map.png)

Based on Google Maps, the elevation is around 6000 feet, which is consistent with the elevation listed in the database, 1838 meters.

# 6. Calculate summary statistics. 
```{r}
elev <- met[which(met$elev == max(met$elev, na.rm = TRUE)), ]
summary(elev)
```
```{r}
cor(elev$temp, elev$wind.sp, use="complete")
```
```{r}
cor(elev$temp, elev$hour, use="complete")
```

```{r}
cor(elev$wind.sp, elev$day, use="complete")
```

```{r}
cor(elev$wind.sp, elev$hour, use="complete")
```

```{r}
cor(elev$temp, elev$day, use="complete")
```

# 7. Exploratory graphs
```{r}
hist(met$elev)
```
```{r}
hist(met$temp)
```
```{r}
hist(met$wind.sp)
```
```{r}
library(leaflet)
leaflet(elev) |>
  addProviderTiles('OpenStreetMap') |> 
  addCircles(lat=~lat,lng=~lon, opacity=1, fillOpacity=1, radius=100)
```

```{r}
library(lubridate)
elev$date <- with(elev, ymd_h(paste(year, month, day, hour, sep= ' ')))
summary(elev$date)
```
```{r}
elev <- elev[order(elev$date), ]
head(elev)
```
```{r}
plot(elev$date, elev$temp, type="l", 
     xlab = "Date", ylab = "Temperature (deg C)", main = "Temperature and Date")
```

It appears that temperature peaks once per day. The range of temperatures remained fairly consistent during the month of August. The temperature typically ranged from 4C to 12C. 

```{r}
plot(elev$date, elev$wind.sp, type="l", 
     xlab = "Date", ylab = "Wind Speed", main = "Wind Speed and Date")
```

It appears that wind speed fluctuates significantly throughout the course of each day, without a single daily peak. The peak wind speed gradually increased from the beginning of August to August 19, and then dipped before peaking again around August 26th. 

# 8. Ask questions

## At the station with the highest elevation, is wind speed related to temperature? 
```{r}
plot(elev$wind.sp, elev$temp, 
    main = "Wind Speed vs Temperature", 
    xlab = "Wind Speed",
    ylab = "Temperature",
    pch = 19, 
    col = "blue")
```

Based on this plot, it appears that wind speed is not related to temperature. The points are randomly distributed with no clear pattern. 

## Is temperature related to relative humidity?
```{r}
plot(met$temp, met$rh, 
    main = "Temperature vs Relative Humidity", 
    xlab = "Temperature",
    ylab = "Relative Humidity",
    pch = 19, 
    col = "blue")
```

As elevation increases, relative humidity decreases. This is expected, because cold air cannot hold as much water vapor as warm air.