Description

This dataset is extracted from the original source $($shown above$)$ to include weather information

observed on July $15,2020$ only. Features of data include

$$ date: month, day, and year when the observation is made

$$ station id: ID of a location where the observation is made

$$ lat: latitude coordinate

$$ lon: longitude coordinate

$$ mean temp: average temperature

$$ wind speed

$$ precipitation

The observations are made from $50$ states across the United States. Our goal is to apply clustering

methods to group locations with similar weather conditions.

Pre$-$processing

$1.$ Import data from weather.csv$,$ save it as myData. Hint: use read.csv$().$

$2.$ Use summary$()$ to check if there are columns with NA values.

$3.$ Remove rows with NA station ID$.$

$4.$ Check if data contains duplicated station ID$.$

$5.$ There are still NA values in columns wind speed and precipitation. We will fill those empty

values using these steps:

$($a$)$ Fill NA precipitation with value $0.$ Hint: is$.$na$()$ returns TRUE$/$FALSE for NA values.

$($b$)$ Fill NA values of wind speed with the national median value.

$($c$)$ Can you think of other ways of filling in the missing values?

$6.$ Remove observations from Alaska and Hawaii.

$7.$ Let us visualize the data. Make a scatter plot using ggplot. Use longitude $($x$-$axis$)$ and

latitude $($y$-$axis$)$ for coordinates, and use column mean temp to color points.

$($Optional$)$ We can change the color$-$gradient by adding

$+$ scale

color

gradient$($low $=$ "color$1",$ high $=$ "color$2"),$

where color$1$ and color$2$ are color names $($e$.$g$.,$ gold, red, etc.$).$

Clustering: We are going to apply clustering algorithms using weather features $($average temper

ature, wind speed, and precipitation$).$

$8.$ Save a subset of myData containing the above $3$ columns only.

$9.$ Apply K$-$means clustering algorithm to group the data points to $5$ clusters.

$($a$)$ Report the number of data points in each cluster.

$($b$)$ For each cluster, report the average temperature, wind speed, and precipitation.

$($c$)$ Repeat Question $7$ to make a scatter plot. This time, we will use cluster membership

$($e$.$g$.,1,2,3,4,5)$ to show different colors. Use as$.$factor$()$ to convert the cluster mem

bership before providing it to ggplot.

$($Optional$)$ We can manually choose colors for each group by adding

$+$scale

color

manual$($values$=$c$("$color$1",$"color$2",$ "color$3",$ "color$4",$ "color$5")).$

$10.$ Apply hierarchical clustering $($use complete link method$)$ on the subset data created in

Question $8.$

$($a$)$ Make a dendrogram.

$($b$)$ Trim the clustering to get $3$ clusters.

$($c$)$ Repeat Question $7$ to visualize the clustering output