- 1. Objective
- 2. Introduction to Data Analytics
- 3. Analysis versus Reporting
- 4. Data Analytics Process
- 5. Introduction to Data Mining
- 6. Characteristics of Big Data Analysis
- 7. Great analysis with framing the Problem Correctly
- 8. Skills required to be a Data Analyst
In this data analytics tutorial, we will learn about what is data analysis, data analytics definition, what are the various business intelligence tools for analyzing data, what is data mining, what is the difference between analysis vs reporting, Introduction to data mining and various data mining techniques, features of Big data analysis, how to do proper analysis by framing the problem statement correctly, what is statistical significance and business importance in terms of business analysis and the various skills required for Big data analyst.
2. Introduction to Data Analytics
Data is information in raw format. With increasing data size, it has become need for inspecting, cleaning, transforming, and modeling data with the goal of finding useful information, making conclusions, and supporting decision making. This process is known as data analysis.
Data mining is a particular data analysis technique where modeling and knowledge discovery for predictive rather than purely descriptive purposes is focussed. Business intelligence covers data analysis that relies heavily on aggregation, focusing on business information. In statistical applications, some people divide business analytics into descriptive statistics, exploratory data analysis (EDA), and confirmatory data analysis (CDA). EDA focuses on discovering new features in the data and CDA focusses on confirming or falsifying existing hypotheses. Predictive analytics does forecasting or classification by focusing on statistical or structural models while in text analytics, statistical, linguistic and structural techniques are applied to extract and classify information from textual sources, a species of unstructured data. All are varieties of data analysis.
The Big Data wave has changed ways in which industries function. With Big Data has emerged the requirement to implement advanced analytics to it. Now experts can make more accurate and profitable decisions.
In this session, we are going to see characteristics and need of data analysis.
3. Analysis versus Reporting
An analysis is an interactive process of a person tackling a problem, finding the data required to get an answer, analyzing that data, and interpreting the results in order to provide a recommendation for action.
A reporting environment or a business intelligence (BI) environment involves calling and execution of reports. The outputs are then printed in the desired form. Reporting refers to the process of organizing and summarizing data in an easily readable format to communicate important information. Reports help organizations in monitoring different areas of a performance and improving customer satisfaction. In other words, you can consider reporting as the process of converting raw data into useful information, while analysis transforms information into insights.
Let us understand difference between data analysis and data reporting:
- Reporting provides data. A report will show the user what had happened in the past, to avoid inferences and help to get a feel of the data while analysis provides answers for any question or issue.An analysis process takes any steps needed to get the answers to those questions.
- Reporting just provides the data that is asked for while analysis provides the information or the answer that is actually needed.
- Reporting is done in standardized manner while analysis can be customized. There are fixed standard formats for reporting while analysis is done as per the requirement and it is customizable as needed.
- Reporting can be done using a tool and it generally does not involve any person while in analysis, person is required who is doing analysis and who will lead the process. He guides the complete analysis process.
- Reporting is inflexible while analysis is flexible. Reporting provides no or limited context about what’s happening in the data and hence is inflexible while analysis emphasizes data points that are significant, unique, or special, and it explains why they are important to the business.
4. Data Analytics Process
Now we are going to see the analytic process or how analyzing data can be done?
a. Business Understanding
The very first step consists of business understanding. Whenever any requirement occurs, firstly we need to determine business objective, assess the situation, determine data mining goals and then produce the project plan as per the requirement. Business objectives are defined in this phase.
b. Data Exploration
Second step consists of Data understanding. For further process, we need to gather initial data, describe and explore the data and verify data quality to ensure it contains the data we require. Data collected from the various sources is described in terms of its application and need for the project in this phase. This is also known as data exploration. This is necessary to verify the quality of data collected.
c. Data Preparation
Next come Data preparation. From the data collected in last step, we need to select data as per the need, clean it, construct it to get useful information and then integrate it all. Finally we need to format the data to get appropriate data. Data is selected, cleaned, and integrated in the format finalized for the analysis in this phase.
d. Data Modeling
Once data is gathered, we need to do data modeling. For this, we need to select modeling technique, generate test design, build model and assess the model built. Data model is build to analyze relationships between various selected objects in the data, test cases are built for assessing the model and model is tested and implemented on the data in this phase.
e. Data Evaluation
Next come data evaluation where we evaluate the results generated in last step, review the scope of error and determine next steps that need to be performed. Results of the test cases are evaluated and reviewed for the scope of error in this phase.
Final step in analytic process is deployment. Here we need to plan the deployment and monitoring and maintenance, we need to produce final report and review the project. Results of the analysis are deployed in this phase. This is also known as reviewing of the project.
The complete above process is known as business analytics process.
5. Introduction to Data Mining
Data mining, also called as data or knowledge discovery, means analyzing data from different perspectives and summarizing it into useful information – information that can be used to take important decisions. It is the technique of exploring, analyzing, and detecting patterns in large amounts of data. Goal of data mining is either data classification or data prediction. In classification, data is sorted into groups while in prediction, value of a continuous variable is predicted.
In today’s world, data mining is been used in several sectors like Retail, sales analytics, Financial, Communication, Marketing Organizations etc. For example, a marketer may want to find who did and did not respond to a promotion. In prediction, the idea is to predict the value of a continuous (ie non-discrete) variable; for example, a marketer may be interested in finding who will respond to a promotion.
Some examples of Data Mining are:
a. Classification of trees
These are Tree-shaped structures that represent sets of decisions.
b. Logistic regression
It predicts the probability of an outcome that can only have two values.
c. Neural networks
These are non-linear predictive models that resemble biological neural networks in structure and learn through training.
d. Clustering techniques like the K-nearest neighbors
This is the technique that classifies each record in a dataset based on a combination of the classes of the k record(s) most similar to it in a historical dataset (where k 1). Sometimes it is called the k-nearest neighbor technique.
e. Anomaly detection
It is the identification of items, events or observations which do not conform to an expected pattern or other items in a dataset.
6. Characteristics of Big Data Analysis
We have already seen characteristics of Big Data like volume, velocity and variety. Let us now see characteristics of Big Data Analytics which make it different from traditional kind of analysis.
Big Data analysis has the following characteristics:
There might be need to write program for data analysis by using code to manipulate it or do any kind of exploration because of the scale of the data.
b. Data driven
It means progress in an activity is compelled by data and program statements describe the data to be matched and the processing required rather than defining a sequence of steps to be taken. Many analysts use hypothesis driven approach to data analysis, Big Data can use the massive amount of data to drive the analysis.
c. Attributes usage
For proper and accurate analysis of data, it can use lot of attributes. In the past, analysts dealt with hundreds of attributes or characteristics of the data source, with Big Data there are now thousands of attributes and millions of observations.
As whole data is broken into samples and samples are then analyzed, data analytics can be iterative in nature. More compute power enables iteration of the models until Big Data analysts are satisfied. This has led to development of new applications designed for addressing analysis requirements and time frames.
7. Great analysis with framing the Problem Correctly
In order to have a great analysis, it is necessary to ask the right question, gather the right data to address it, and design the right analysis to answer the question. Then only analysis can be called as correct and successful.
Framing of problem means ensuring that important questions have been asked and critical assumptions have been laid out. For example, is the goal of a new initiative to drive more revenue or more profit? The choice leads to a huge difference in the analysis and actions that follow. Is all the data required available, or is it necessary to collect some more data? Without framing the problem, the rest of the work is useless.
For great analysis, problem should be framed correctly. This includes assessing the data correctly, developing a solid analysis plan, and taking into account the various technical and practical considerations in play.
Any business problem can be analyzed for 2 issues:
a. Statistical Significance
How problem is statistically important for decision making. Statistical significance testing takes some assumptions and determines the probability of happening of results if the assumptions are correct.
b. Business Importance
It means how the problem is related with business and its importance. Always put the results in business context as part of the final validation process.
8. Skills required to be a Data Analyst
In today’s world, there is an increasing demand for analytical professionals. It is taking time for academic programs to adapt and scale to develop more talent.
All the data collected and the models created are of no use if the organization lacks skilled Big Data analysts. A Big Data analyst requires both skill and knowledge for getting good data analytics jobs.
To be a successful analyst, a professional requires expertise on the various Big data analytical tools like R & SAS. He should be able to use these business analytics tools properly and gather required details. He should also be able to take decisions which are both statistically significant and important to the business.
Even if you know how to use a data analysis tool of any type, you also need to have the right skills, experience and perspective to use it. An analytics tool may save a user some programming but he or she still needs to understand the analytics that are being generated. Then only a person can be called as successful Data analyst.
Business people with no analytical expertise may want to leverage analytics, but they do not need to do the actual heavy lifting. The job of the analytics team is to enable business people to drive analytics through the organization. Let business people spend their time selling the power of analytics upstream and changing the business processes they manage to make use of analytics. If analytics teams do what they do best and business teams do what they do best, it will be a winning combination.