Data analysis and visualization skills

In this section, I put some of my R, SQL and Java code that I have used in various classes at George Mason University.

R programming:

Quality Management (OSCM 456)

library(ggplot2)

CACI_rev <- data.frame (Year = c(2018, 2019, 2020), Revenue = c(4467860, 4986341, 5720042))

CACI_op_income <- df <- data.frame (Year = c(2018, 2019, 2020), Income_from_operations = c(340700, 377867, 457696))

CACI_income <- df <- data.frame (Year = c(2018, 2019, 2020), Net_Income = c(301171, 265604, 321480))

ggplot(

CACI_rev, aes(

    x = Year, y = Revenue)) + geom_line(col = "purple", size = 2) + geom_point(col = "blue", size = 8) + ylim(3000000, 6000000) + labs(title = "Revenue from 2018 to 2020", y = "Revenue", subtitle = "The historical revenue of CACI form 2018 to 2020") + theme_minimal()

ggplot(

CACI_op_income, aes(

    x = Year, y = Income_from_operations)) + geom_line(col = "cyan", size = 2) + geom_point(col = "blue", size = 8) + ylim(250000, 500000) + labs(title = "Operating Income from 2018 to 2020", y ="Operating Income", subtitle = "The historical Operating Income of CACI form 2018 to 2020") + theme_minimal()

ggplot(

CACI_income, aes(

    x = Year, y = Net_Income)) + geom_line(col = "green", size = 2) + geom_point(col ="blue", size = 8) + ylim(200000, 400000) + labs(title = "Net Income from 2018 to 2020", y = "Net Income", subtitle = "The historical Net Income of CACI form 2018 to 2020") + theme_minimal()

Plotting:

Neural Networks in R:

Advanced Data Mining (MIS 432)

library(neuralnet)

employment_df <- read.csv("Project_pre_processed.csv", header = TRUE)

#Review the dataset

View(employment_df)

#The dataset includes the observation value, delete it

employment_df <- employment_df[, -1]

#The employment change will be the variable that we want to predict using the data

#Divide the dataset into training and validation

set.seed(1)

train_rows_e <- sample(rownames(employment_df), dim(employment_df)[1]*0.6)

train_data_e <- employment_df[train_rows_e, ]

valid_rows_e <- setdiff(rownames(employment_df), train_rows_e)

valid_data_e <- employment_df[valid_rows_e, ]

#Create the model

set.seed(1)

nn_employment_22 <- neuralnet(Employment_Change_Scaled ~

Education.Code + trCode + Med_Ann_Scaled + X5_or_more + X5_or_less, data = train_data_e, linear.output = TRUE, hidden = c(2, 2))

plot(nn_employment_22)

str(nn_employment_22)

#Responses are numeric, use MSE

#First, predict the results for the model

nn_22_pred <- predict(nn_employment_22, valid_data_e[, c(1, 2, 3, 7, 8)])

(MSE_nn_22 <- sum((valid_data_e$Employment_Change_Scaled - nn_22_pred)^ 2)/ nrow(valid_data_e))

#Really large datasets may make the computation time to be longer, not the case

#Test theory of partition, try an 80/20 partition instead

set.seed(1)

train_rows_e8 <- sample(rownames(employment_df), dim(employment_df)[1]*0.8)

train_data_e8 <- employment_df[train_rows_e, ]

valid_rows_e8 <- setdiff(rownames(employment_df), train_rows_e)

valid_data_e8 <- employment_df[valid_rows_e, ]

#Do the other neural networks

set.seed(1)

nn_employment_3_8 <- neuralnet(Employment_Change_Scaled ~

Education.Code + trCode + Med_Ann_Scaled + X5_or_more + X5_or_less, data = train_data_e8, linear.output = TRUE, hidden = 3)

plot(nn_employment_3_8)

str(nn_employment_3_8)

set.seed(1)

nn_employment_22_8 <- neuralnet(Employment_Change_Scaled ~

Education.Code + trCode + Med_Ann_Scaled + X5_or_more + X5_or_less, data = train_data_e8, linear.output = TRUE, hidden = c(2, 2))

plot(nn_employment_22_8)

str(nn_employment_22_8)

K-means clustering using the tidyverse:

Advanced Data Mining (MIS 432)

#Based on https://www.tidymodels.org/learn/statistics/k-means/

#A tidy approach to K means clustering

library(tidyverse)

library(tidymodels)

#Convert the data frame into a tibble

utilities_df <- read.csv("Utilities.csv", header = TRUE)

#Apply all the transformations

uti_tibble_norm <- as_tibble(utilities_df) %>% column_to_rownames(var = "Company") %>% modify(scale)

View(uti_tibble_norm)

set.seed(5)

(util_tidy_clust <- kmeans(uti_tibble_norm, centers = 6))

#Let's explore the model first with summary

summary(util_tidy_clust)

#We can see the cluster classification with augment from tune

augment(util_tidy_clust, uti_tibble_norm) %>% select(.rownames, .cluster) %>% print(n = 22)

#Tidy will summarize by cluster

tidy(util_tidy_clust)

#Glance gives us access to a single-row summary

glance(util_tidy_clust)

SQL code:

Data Warehousing (MIS 430)

/* Fact Tables */

CREATE TABLE [Avg_Price Fact Table] (

[LocationKey] int NOT NULL,

[CalendarKey] int NOT NULL,

[VolumeKey] int NOT NULL,

[CompanyKey] int NOT NULL,

[ClosingPrice] float NOT NULL,

[OpeningPrice] float NOT NULL,

[Average_Price] float NOT NULL

PRIMARY KEY(LocationKey, CalendarKey, VolumeKey, CompanyKey)

)

CREATE TABLE [Avg_Volume Fact Table] (

[CalendarKey] int NOT NULL,

[CompanyKey] int NOT NULL,

[VolumeKey] int NOT NULL,

[VolumePrice] float NOT NULL

)

/*Dimensions*/

CREATE TABLE [Volume Dimension] (

[VolumeKey] int NOT NULL PRIMARY KEY,

[VolumeCategory] char,

[DatasetKey] int NOT NULL,

)

CREATE TABLE [Company Dimension] (

[CompanyKey] int NOT NULL PRIMARY KEY,

[CompanySectorName] char NOT NULL,

[CompanyName] char NOT NULL,

[CompanySymbol] char NOT NULL

)

CREATE TABLE [Location Dimension] (

[LocationKey] int NOT NULL PRIMARY KEY,

[State] char,

[City] char,

[Street] char,

[Zip] char

)

CREATE TABLE [Calendar Dimension] (

[CalendarKey] int NOT NULL PRIMARY KEY,

[FullDate] date,

[Year] int,

[Month] char,

[Quarter] char,

[DayofWeek] char

)

Java Programming:

Programming for Business (MIS 350)

1 import java.awt.*;
2 import java.awt.event.*;
3 import javax.swing.*;
4 public class ECTechRepair extends JFrame implements ActionListener {
5 JPanel panel;
6 JLabel user_label, password_label, message;
7 JTextField userName_text;
8 JPasswordField password_text;
9 JButton submit, cancel;
10 ECTechRepair(){
11 //This is the Username Label
12 user_label = new JLabel();
13 user_label.setText("Employee ID");
14 user_label.setOpaque(true);
15 user_label.setBackground(Color.YELLOW);
16 Font font = new Font("Times New Roman", Font.BOLD, 25);
17 user_label.setFont(font);
18 userName_text = new JTextField();
19 //Label for the Password
20 password_label = new JLabel();
21 password_label.setText("Employee Password :");
22 password_label.setOpaque(true);
23 password_label.setBackground(Color.YELLOW);
24 password_text = new JPasswordField();
25 password_label.setFont(font);
26 //Enter
27 submit = new JButton("Enter Credentials");
28 submit.setBackground(Color.CYAN);
29 panel = new JPanel(new GridLayout(3, 2));
30 panel.setBackground(Color.CYAN);
31 panel.add(user_label);
32 panel.add(userName_text);
33 panel.add(password_label);
34 panel.add(password_text);
35 message = new JLabel();
36 panel.add(message);
37 panel.add(submit);
38 setDefaultCloseOperation(JFrame.EXIT_ON_CLOSE);
39 //Add the listeners to the components
40 submit.addActionListener(this);
41 add(panel, BorderLayout.CENTER);
42 setTitle("ECTechRepair Unique Login");
43 setSize(600,325);
44 setVisible(true);
45 }
46 public static void main(String[] args){
47 new ECTechRepair();
48 }
49 public void actionPerformed(ActionEvent ae){
50 String EmployeeID = userName_text.getText();
51 String password = password_text.getText();
52 Font font = new Font("Times New Roman", Font.BOLD, 35);
53 if (EmployeeID.trim().equals("143") && password.trim().equals("Demo")) {
54 message.setText(" Hello Fariha");
55 message.setFont(font);
56 message.setForeground(Color.BLUE);
57 } else {
58 message.setText("Invalid credentials");
59 message.setFont(font);
60 message.setForeground(Color.RED);
61 JOptionPane.showMessageDialog(null, "Try again");
62 }
63 }
64 }