Big Data Cheat Sheet

Big data involves sifting through large amounts of data to find meaning or something of use. Data drives the world economy because the world is increasingly becoming automated. Take the stock market, for example; about 80% of the stock market is automated. So much data is relied on every single data, from what you Google to what you buy on the internet. Corporations use data from products and services to find trends, statistics, and facts that could help improve customer satisfaction and earn a profit. But it’s not the vast amount of data that is received daily that is important, but the insight gained from the data.

This is where a comes in handy. The certification will teach data analytics, infrastructure and tools used by big data engineers and other professionals that help you become even more competent in your career. It shows you have the skills and knowledge but also the initiative to go out of your way to earn a certification that expresses credibility in your field.

What Is Big Data?

Big data is collecting, analyzing and finding meaningful insight in the data to make intelligent decisions. Although the idea of big data was formed in the early 2000s by Doug Laney, the big data industry is still evolving and adapting, just as the technology sectors continues to grow at a rapid pace and open up new fields for young professionals. Big data could involve reducing time and cost on a product, finding trends in popular products during certain seasons and capitalizing on what is working to capture more market share in a sector. It could also mean more specific things, such as preventing security breaches at your company and building a strategy around what to do if a data breach happens. The importance of big data can’t be overstated, as it’s necessary to a healthy business no matter if it’s a major corporation or a start-up out of your garage. If you want to grow, you understand that understanding insights from the data is essential. 

Every sector uses big data is come way. Whether it’s retail, healthcare, manufacturing or education, big data helps them create a strategy, understand where the data comes from, manage and store the data, gather insight and use what is learned in an intelligent manner.

Data Science Essentials

Like our course, Data Science Essentials, data science involves R programming, Python, Azure machine learning and various other programming languages. Learning data science also involves becoming familiar in C++, SQL and Java. Eventually, mastering these basics forms a solid foundation upon which you use to move up into positions with higher authority and pay. Let’s go over each one and what to know for tests or interviews involving each:

Start Your 30-Day FREE TRIAL with Data Science Academy to Launch Your Career in Big Data.

R Programming Facts

R Programming: This is a programming language designed for statistical analysis, and you can express your results through graphs and other graphics that explain your results to yourself and to others.

There is a total of six data objects in R Programming: lists, arrays, vectors, matrices, tables, and data frames.

To use a csv file in R: Use the read.csv function. It uses a data frame to help you access the file.

What’s a valid variable name? This is a name that consists of numbers, letters, and dotted or underlined characters.

The difference between array and matrix: A matrix is two-dimensional with rows and columns. An array can have many dimensions, and a matrix is one dimension.

What stores and processes categorical data? These are factor data objects.

How to find the name of the current working directory? This is through the command getwd()

R Base package: This is the default package loaded when the R environment is set. The basics such as math calculations and inputs/outputs are found here.

How is R used in logistic regression? You use glm() to use logistic regression, and it’s used for the probability of a binary response variable.Create a boxplot graph: Use boxplot()

How to find all the packages? Use the command: installed.packages()

What is inlist()? This turns a list to a vector.

Give the R regression for when you toss a coin 51 times using pbinom and get 26 or less heads. 

x <- pbinom(26,51,0.5)

print(x)

What is as.data.frame()? This converts data in a JSON file to a data frame.

Create a histogram: hist()

Remove vector from R workspace: rm(x)

List all data sets in the available packages:

data(package = .packages(all.available = TRUE))

Create scatterplot matrices: pairs(formula, data)

SQL-Related Information

SQL is a programming language that’s useful for big data, as it is one way of pulling information from databases and making some use for the data. It’s an old language but still common among programmers working in the big data field. You can query and reorganize data while also updating the structure of the database. An SQL database could store data from a customer and go back years to extract data you want from said customer, such as the cost of the service/product, names, amounts, problems encountered and so on. A big data architect uses SQL to build products that handle a huge amount of data. A database migration engineer, data scientist and database administrator all make use of their SQL skills. Even though it’s a fairly easy language to learn, it is still used today for its ability to handle databases. Advanced level SQL is used by the positions listed above.

The following includes facts, questions and answers to SQL-related facts you might be asked about in an interview or test:

What is SQL? Structured Query Language (SQL) involves extracting information from a database. It’s a basic language for databases. You can use it for retrieving, updating, inserting and deleting data from a database.

Query used to find the employee with the highest salary on record:

Select * from table_name where salary = (select max(salary) from table_name);
For example
Select * from employee where salary = (select max(salary) from employee);

Query to find employee with second-highest salary:

There are a few ways of answering. Here is one possible way of answering:

Select Salary from employee where salary in (select salary from employee where level = &topnth connect by prior Salary > Salary group by level)

The value of topnth, in this question, is 2 (for 2^nd highest salary). If the interviewer asks for 3^rd highest salary or 4^th highest, it would be 3 or 4, depending on the question.

Query to find 2^nd lowest salaried employee in an employee table:

Select Salary from employee where salary in (select salary from employee where level = &lownth connect by prior Salary < Salary group by level)

(Change lownth if the question is about 3^rd lowest or 4^th lowest, etc. In either case, it would change for 2 to 3 or 4.)

What is NVL and NVL2? What’s the difference between the two?

Both change a NULL value to an actual one. There are two parameters for the NVL function, but the NVL parameter has three arguments.

NVL (expr1, expr2)

expr1 is the source value containing the NULL

expr2 is the target value that will convert the NULL

C++

C++ is another common programming language used in big data. It’s an object-oriented programming language that is praised for its fast speed. Over 1GB of data can be processed with one second, which makes it excellent for managing vast amount of data. It is also used for deep learning algorithms and for machine learning. Machine learning will be explained later, along with keywords and questions that might be asked during an interview or test. C++ requires less power and capacity, which reduces costs. Since it’s so powerful, fast and useful, it’s essential to big data. The following includes questions you might see during an interview or test involving C++:

What is C++? An object-oriented programming language, a superset of C, used for its vast variety of features.

Is C++ part of OOPS? Yes. OOPS is object-oriented programming system and involves a paradigm providing applications concepts such as inheritance, polymorphism, data binding and more.

What is class in C++? Class involves designing user-defined data types. It differentiates attributes, entities and actions.

What is encapsulation? It’s the process of binding data and functions together in a class. For security reasons, it is used to prevent direct access to the data, similar to how a bank asks for a login id and password before entering the portal.

Define abstraction. The process involving hiding internal implementations and showing only the essential details. An example is a success message after sending an email, yet you aren’t told about the process of the email as it is sent since that’s irrelevant; only whether the email was sent successfully or not is relevant.

What does the keyword “volatile” do? This function helps you determine that a certain variable is volatile, then the compiler is directed to change the variable.

What is a storage class? Storage class is an assortment of symbols to separate variables, functions, etc. Various storage class names include extern, auto, static, register, mutable, etc.

Is a recursive inline function possible in C++? A compiler with great optimization could perform inline recursive calls. But if there isn’t good optimization, it won’t be able to find the recursion depth at the compile time.

What is an inline function? They reduce function call overhead. As the name implies, the function expands once it’s called. The syntax is found below:

Inline return-type function-name(parameters)

{

// Function code goes here

}

What is “this” pointer? It captures the address of the memory of the current object. It signals through a hidden argument to the collective non-static function calls. It can also be used as a local variable inside the body of all non-static functions.

Why is the Friend class and function needed? It’s needed to access protected or private members of a class. The friend class can access protected and private members that are labeled as a friend. A friend function is either a global function or a method as part of a class. Friendship isn’t inherited or mutual, such as a class wanting to be a friend to another class that labels it ‘NotAFriend’ cannot automatically become its friend. 

What’s the difference between function overloading and operator overloading? Operator overloading enables redefining how an operator works for user-defined types. Function overloading enables two or more functions with a different amount of parameters and types to have the same name.

Compare C++ with Java. C++ includes destructors that begin when an object is destroyed. But Java has an automatic garbage collection instead. C++ supports structures, templates, inheritance, pointers, unions and operator overloading; Java does not include these. C++ doesn’t have an inbuilt support for threads, but Java has a thread class used to create a new thread. C++ uses a compiler to translate the source code into a machine level language, which means it is platform-dependent. Java isn’t platform dependent. Java converts the code into JVM bytecode.

What are the main differences between C and C++? C doesn’t support references; C++ does. C++ includes inheritance, templates, friend functions, virtual functions and function overloading. But these are not found in C. C++ deals with exception handling at the language level, but uses the common if-else style for exception handling. C++ supports procedural and object-oriented programming approaches, and C only uses a procedural programming language.

What is a static member? The default access specifier is public when deriving a base class or structure while it’s private when deriving a class. By default, members of a structure are public while members of a class are private.

What is a destructor? Can you overload it? You can’t overload it, and it’s a member function of a class. It’s used automatically when an object eludes its scope.

Differentiate keyword struct and class. The keyword struct is utilized when you want to resemble public members by default, and keyword class for when resembling private members by default.

Define namespace. This resolves a naming conflict by the identifier. This is solved by placing these under various namespaces.

What’s a class template? It’s the name of a generic class. The keyword template is used to define a class template

What is a token in C++? It’s a name given to many functions within C++. This can be a keyword, constant, identifier, symbol, string literal and more.

This only touches on C++, since, like any programming language, you can’t go over everything all on one sheet (unless that cheat sheet goes on for thousands of pages). Some websites contain a lot of essentials that may be of use when looking for basic interview questions and answers or simply reviewing facts about Python, machine learning, artificial intelligence, Hadoop, C++ and much more.

General Interview Questions Relating to Big Data Professionals

Could you describe big data for me? Assume I know nothing about it.

Big data deals with large and complex databases. You can sift through large amounts of data using tools, and from the information gain insight to make better business decisions.

Name the five v’s of big data.

Velocity, variety, volume, veracity, value. Value adds some usefulness to the business, such as by helping them gain more revenue. Volume represents the amount of data that increases all the time. Velocity is how fast the data increases. Variety is the different kinds of data. Veracity involves the uncertainty from receiving so much data every day.

How does analyzing big data increase revenue in a business?

One example is through going through the data and witnessing trends and capitalizing on these trends. Another is differentiating itself by looking at competitors, finding out what customers want and need, and using this information to create something new and useful to the customer.

What are the three steps to a big data solution?

Data ingestion, data storage, and data processing.

What’s fsck?

File System Check is an HDFS command. It’s used to check for problems. If there’s something missing, for example, it’ll find it.

Now we’ll go over basic questions relating to Hadoop since Hadoop is commonly used by big data professionals.

Name some popular input formats in Hadoop.

Key-value input format, text input format, and sequence file input format.

What are the core components of Hadoop?

• HDFS – Hadoop Distributed File System is the foundation by which large data files are stored. It can store data when the hardware fails.
• YARN – This is a cluster resource management that includes several data processing engines.
• MapReduce – This deals with data processing. The two phases are a map and reduce. Map details complex logic code and Reduce handles light-weight operations.

Name the three running modes.

• Standalone/local – This is the default mode. NameNode, DataNode, ResourceManager and NodeManager are the components that run of a single JVM.
• Pseudo-distributed – A single node is used to display Hadoop master and slave services.
• Fully-distributed – Separate nodes are used to deploy and execute Hadoop master and slave services.

What are some key features in Hadoop?

• Reliable – Information is stored independently so losing data isn’t a worry.
• Scalable – It’s easily compatible with other systems.
• Open source – It’s open-source and can be changed.
• Distributed processing – Faster processing due to distributed processing.

Differentiate Hadoop and RDBMS

• Hadoop has an open-source framework while RDBMS is licensed software.
• Hadoop is based on Schema on Read and RDBMS is based on Schema on Write.
• Hadoop has semi-structured, structured, and unstructured data types. RDBMS only has structured data.
• Hadoop applications include data discovery and storage and processing of unstructured data. RDBMS has OLTP and complex ACID transactions.

The other questions that may pertain to you depend on if you have experience or not. If so, they’ll ask about the experience using Python or C++ and what you did for your company. If not, they may ask basic questions about the above features relating to Hadoop and other common tools that big data professional use. Not every listed question on websites pertain to each person, and not every business will use the same tools and programming languages. Yet, looking up big data jobs on Indeed, you can see many of them ask about Hadoop (or other big data technologies like Spark, Presto, Kafka), Java or Python experience, SQL, Amazon Web Services, and DevOps methodologies. It takes time to learn much of what is asked by employers.

Artificial Intelligence (AI)

• What is artificial intelligence? An area of computer science relating to computer software making intelligent decisions, reasoning, and problem-solving.
• What is an algorithm? A set of rules used to perform a task. The algorithm informs the machine how it should go about finding the answers to a question or issue.
• What are artificial neural networks? These are learning models based on neural networks found in animal brains. They are used to solve complex tasks that traditional ways could not solve. 
• Define data mining. Patterns found in massive data sets that are found in order to pull useful information.
• What is deep learning? Part of machine learning, deep learning uses special algorithms to comprehend complex data structures and relationships within data and datasets
• Define machine learning. This is a field of AI involving machines that act without being told to do so. Machines learn over time what to do and adjust based on past results.

Data Science Academy's Big Data Courses

Data Science Academy has over 30 years of experience in the eLearning industry and is accredited by Cisco, Microsoft, NetApp, CompTIA, ITIL, and many other certification providers. We have courses covering data science, cloud computing, DevOps, information security, IT ops, application development, and more.

We hope you found this big data guide useful. Earning a big data certification could improve your chances of advancing in your career. If you are interested in Data Science Academy, it has over 30 courses related to big data, including topics such as R programming, machine learning, Transact SQL, Hadoop, artificial intelligence, and more.