You have been asked by management (manufacturing, healthcare, retail, financial, etc. ) to create a demo using a data analytic or BI tool.

Running Head: MIDTERM PROJECT 1

MIDTERM PROJECT 7

Midterm Project

Introduction

	This project is an analysis of a demo financial dataset that my manager needs to understand big data analytics capabilities. In this case, we used the Financial Samples dataset and Microsoft Power BI tools to understand how different products work in the US market and explain the history of Microsoft Power BI and its benefits in the process. And the disadvantages, the type of data I'm looking at, how to browse the data using Power BI tools, etc. Analyze your data using Power BI tools, using classification basics and decision trees. Later, I will briefly introduce alternative classification methods such as rule-based methods, memory-based inference, and neural networks, and summarize the results.

The purpose of this research effort is to use big data analytics and use data analytics tools such as Microsoft Power BI, R Studio, Tableau, and other data analytics tools to understand the concept of decision trees and classifiers. is. Various data mining techniques are used for analysis.

History of Tool

The Microsoft Power BI project was started as a secret Microsoft task in 2006, codenamed "Gemini" by Amir Nets, and later inherited from SQL Server Analysis Services (SSAS). In 2009, it was renamed to "Power Pivot" and released as an Excel extension. Power Pivot was increasingly recognized from 2010 to 2012 and was released in the form of SharePoint's "PowerView" to challenge traditional visual platforms like Tableau. But it's not very popular. Microsoft also released the Excel add-in "PowerQuery" in 2013. However, there was a problem with the automatic data scheduler after the upgrade and there was no central user interface. Finally, after several failed trials and prototyping, Microsoft announced the Power BI platform in 2015. It quickly caught the attention of people with its powerful dashboard, reporting system, relationship platform and convenient user interface. (Rodman, 2018).

According to Tim Rodman (2018), "Power BI: A tool that can turn a" data cow "into a delicious and easy-to-digest" data steak. " However, like everything else, Power BI has its strengths and weaknesses. .. Less than:

Advantages of Microsoft Power BI

• Microsoft Power BI integrates seamlessly with existing business environments and datasets (SQL files, Excel datasets, etc.) and can also enhance analysts' reporting and analytics capabilities (Mane, 2018).

• Custom dashboards are very easy to use and powerful, but Power BI makes it easy for users to view reports (Mane, 2018).

• Publish reports easily and securely with automatic data update options that keep you up to date with the latest version of information (Mane, 2018).

• Cloud system changes allow you to quickly restore data without memory and speed limits (Mane, 2018).

• Power BI has the least knowledge of data analysis and knows how to work with the Internet domain, so anyone can use Power BI (Mane, 2018).

• Power BI integrates simplicity and performance at the same time (Mane, 2018).

• Power BI can support advanced cloud data services such as Cortana and Bot frameworks (Mane, 2018).

• Extract business intelligence insights very quickly and accurately to help business managers improve decision-making efficiency.

Limitations of Microsoft Power BI

• The dynamic capabilities of Customer Relationship Management (CRM) make Power BI very efficient, but there are security issues. All Power BI tool users can use the data model (Schaeffer, nd) unless the author grants certain security privileges.

• Dashboards and reports can be easily shared with Office 365 tenants with the same email domain (Schaeffer, n.d.).

• Power BI can't accept files larger than 250MB. This is the main drawback of this data warehousing era (Schaeffer, n.d.).

• Each dataset is limited to 1GB. You may need to process multiple datasets with the same data. This limits functionality (Schaeffer, n.d.).

• Microsoft Power BI is still in the prototype stage and needs further development (Schaeffer, n.d.).

Review of the Data

I checked the sample demo data. The file name is "Financial Sample". Collected as a sample dataset from the whiteboard platform. In these datasets, sales and revenues for Carretera, Montana, Paseo, VTT, Amarilla, and Velo were analyzed based on five countries in two different regions (North America and Europe). These countries are the United States, Canada, Mexico, France and Germany. We left some columns in the dataset, but split the dataset for analysis to get data on sales, country / region, total sales, products, profits, breakdown, unit sales, and year. Did. .. Splitting datasets helps you better visualize what you need to understand from these datasets.

Exploring the Data with Microsoft Power BI

Power BI is very easy and easy to use and will guide you on the right path. First, after opening the Power BI platform, I uploaded the "Financial Sample" dataset via the "Get Data" query system. After loading the data, I edited the dataset with "Edit Query" and deleted unnecessary columns. If you need to find more relationships between your data, you can still get the dataset. To create a decision tree, I clicked the From Marketplace button and got the TreeViz from Power BI Custom Visuals. This helps to display the image as a decision tree. For TreeViz, 1. There are two new segments, categorical data and 2. measurement data. You entered the country / region, product, year, and segment into the categorical data based on the categorical data used in the analysis. Next, put sales, total sales, profit, and unit sales into the measurement data segment. as a result, Two images were obtained. Maps of 1.5 countries and 2. Total. It contains information that has been filtered based on the measurement data (Microsoft, n.d.).

Symmetric encryption:

Symmetric encryption is said to be the simplest and most famous encryption technology. As mentioned above, use the key for encryption and decryption.

* This is the recommended technology for batch transfer of data, as the algorithms behind symmetric encryption are less complex and run faster.

* Plain text is encrypted with a key, and the receiver uses the same key to decrypt the received encrypted text. The host of the communication process receives the key via external means.

* Widely used SE algorithms are AES-128, AES-192, and AES-256.

Asymmetric encryption:

Compared to symmetric cryptography, this type of cryptography is relatively new and is also known as public key cryptography.

* Asymmetric encryption is more secure than symmetric encryption because it uses two keys for processing.

* Anyone can use the public key used for encryption, but the private key will not be disclosed.

* This asymmetric encryption method is used for daily communication on the Internet.

* If you use your public key to encrypt your email, you can only use your private key to decrypt your email. However, if you use the private key to encrypt the message, you can use the public key to decrypt the message.

* You can use the client-server model digital certificate to discover the public key.

* The disadvantage of this encryption is that it takes longer than the symmetric encryption process.

* The most usual asymmetric encryption technologies include RSA, DSA, and PKCS.

Key Difference:

* Symmetric encryption is an old technology, but asymmetric encryption is the latest technology.

* Due to the complexity of common sense, performing asymmetric encryption takes time. For this reason, symmetric encryption is used when transferring records in bulk.

* Asymmetric data is more secure because it uses a unique key for encryption and decryption.

Determine which is the safest.

Asymmetric data is more secure because it uses different keys between encryption and decryption.

Reason:

Asymmetric encryption is more secure than symmetric encryption because it uses two keys to process. Anyone can use the public key for encryption, but the private key is not exposed.

After encrypting the message with the public key, only the private key can be used to decrypt the message

Symmetric encryption

Let's say Alice wants to talk to Bob. He wants to keep the news secret. Bob is the only person who should be able to read the message. Since the message is sensitive, Alice uses the key to encrypt the message. The original message is called plaintext, and the encrypted message is called encrypted text. The ciphertext is sent to Bob, who knows the key and uses the same symmetric encryption (such as AES or 3DES). Therefore, Bob can decrypt the message.

Alice and Bob share a key called the symmetric key. They are the only ones who know the key and cannot read the encrypted message. In this way, confidentiality is achieved.

Asymmetric encryption

Asymmetric encryption uses two keys, a public key and a private key (such as RSA). Anyone can use the public, but only the owner knows about the private. After the message is encrypted with the public key, only the corresponding private key can decrypt the message. Also, you cannot learn the private key from the public key.

Asymmetric encryption solves the problem of secure key distribution. Alice gets Bob's public key and uses it to encrypt the session key. Only Bob knows the corresponding private key, so only Bob can decrypt the encrypted session key. Asymmetric ciphers are only used for secure key distribution because they are much slower to execute than symmetric ciphers. Alice and Bob can then use symmetric encryption and a session key to keep the communication secret.

Asymmetric encryption also solves scalability issues. Everyone needs only a public and private key to communicate with others.

Reason:

Asymmetric encryption is more secure than symmetric encryption because it uses two keys to process. Anyone can use the public key for encryption, but the private key is not exposed. After encrypting the message with the public key, only the private key can be used to decrypt the message

Classifications Basic Concepts and Decision Trees

According to Tan, Steinbach, and Kumar (2005), classification is "a task of assigning objects to one of several predefined categories and is a common problem covering many different applications" (p. 145). ). How to detect spam, how to classify cells according to the quality of MRI scans, etc. (Tan, Steinbach and Kumar, 2005). Similarly, the taxonomy is "a collection of records (training sets), where each record contains a set of attributes, one of which is a class" (Tan, Steinbach and Kumar, 2005, p.146). In classification, the objective function is called the "classification model" and is useful in the following situations: Descriptive modeling. Predictive modeling, etc. (Tan et al., 2005).

Decision Trees

A decision tree is a "hierarchical structure consisting of nodes and directed edges" (Tan et al., 2005, p.150). There are three types of nodes, such as "root nodes" with non-zero inner or outer edges. The second is the internal node, which has an edge inward from the root node and has two or more protruding edges. Finally, the "leaf" or "terminal" node has one leading edge, but no back. edge. In the decision tree, each node is assigned a class level, and non-terminal nodes (such as internal nodes and root nodes) contain attribute test conditions to isolate records with different characteristics (Tan et al. , Al., 2005, page 150).

Classifications Alternative Techniques

Decision trees are a very useful classification method. However, there are other alternatives such as rule-based methods, nearest neighbor classifiers, artificial neural networks (ANN), naive Bayes classifiers, and support vector machines. Briefly, the next section describes each object.

First, rule-based classifiers are the classifier techniques used to classify records, such as using a set of "if ... then ..." rules. "The rules of the model are expressed in the separated normal form R = (r1Vr2V ... rk), where R is called the ruleset and ri is the classification or separation rule" (Tan et al., Al, Al, 2005, p. Page 207). The second is the nearest neighbor classifier. This requires three things, such as the stored recordset, the distance metric used to calculate the distance between records, and the value of k, which is the number of nearest neighbors to get (Tan et al. People,2005). The third is an artificial neural network (ANN), which is an assembly model of interconnected nodes and weighted links. Where the output node is the sum of each input value and is a function of the link weights(Tan et al. (Al., 2005). Fourth is the naive Bayes classifier, which believes that if one of the conditional probabilities is zero, then the entire equation is zero. This alternative classifier is for noise points. It helps to isolate irrelevant attributes. It's very useful. The last alternative classification technology I'd like to briefly explain is a support vector machine or SVM. It's very suitable for processing high dimensional data and dimensions. Avoid the curse of the problem. SVM uses a subset of training examples to represent decision limits, called "support vectors" (Tan et al., 2005, p. 256). Others There are classification methods, but my explanation is limited to the five alternative decision tree classifiers above.

Summary of Results

The results can be analyzed according to the required knowledge. For example, I have attached three separate images (visuals 1, 2, and 3) to this document to better understand the viewer. It can generate countless images, but due to time constraints and page limitations, we only collected three. However, I think these screenshots are easy to understand. If you operate as shown in the picture, the total sales, profit, and total sales of the product may differ depending on the country. You can also view based on year, sales, and profit. The decision tree starts from "total" to five different countries. You can then categorize by product to see the internal relationships between them. In addition, it can be sub classed again based on 2013 and 2014. Finally, it can be sub classed according to the type of business (company, government, small business, etc.).

Visual 1

Note: The visual has been generated by Microsoft Power BI tools based on “Financial Sample.xlsx” data sheet.

Visual 2

Note: The visual has been generated by Microsoft Power BI tools based on “Financial Sample.xlsx” data sheet.

Visual 3

Note: The visual has been generated by Microsoft Power BI tools based on “Financial Sample.xlsx” data sheet.

Power BI allows you to share reports on dashboards and track relationships between variables. Power BI is a very useful analytics platform for us. I would like to learn how to use it and its function. This is one of the basic goals of this medium-term project

References

Mane, Priyanka. (2018, July 06). 8 Major benefits of Microsoft Power BI you must know. Retrieved September 23, 2018, from http://www.saviantconsulting.com/blog/8-major-benefits-of-microsoft-power-BI.aspx

Microsoft. (n.d.). Power BI. Retrieved September 23, 2018, from https://docs.microsoft.com/en-us/power-bi/guided-learning/

Rodman, Tim. (2018). What is Power BI? Retrieved September 23, 2018, from https://www.timrodman.com/what-is-power-bi/

Schaeffer, Chuck. (n.d.). Microsoft Power BI Review. Retrieved September 23, 2018, from http://www.crmsearch.com/powerbireview.php

Tan, P., Steinbach, M. & Kumar, V. (2005). Introduction to data mining. Boston: Pearson Addison Wesley.