Data Analysis is important to every organization to survive in this competitive world. In recent years every one wants to make use of the data, understanding their business and to take effective decisions. I spent around 10 years in analyzing the data. And played different roles to help my clients to take effective decisions. In this topic I am going to discuss the basics of data analysis to give an overall idea about Data Analysis.
Data Analysis – Learning Path
Data Analysis – Learning Outcomes
- You will understand – What is data analysis? And why do we need to analyze the data?
- Different kinds of requests and the uses of outputs in Data Analysis
- And Steps and approach to follow to analyse the data
What is Data Analysis
“It is a process of understanding and analyzing the data to draw hidden facts to aid decision making.”
We need to answer many questions to sustain in this competitive world. To do this we use the historical or primary data. And we analyze it to draw the conclusions. We describe the facts and detect patterns to solve the business problems. Data Analysis is not a tool or a software. It is a practice and a way to understand and interpret the data. It helps to derive the value out of your data.
Why do we need Data Analysis?
“The quick answer is derive the insights to drive the business successfully.”
Every day we are (humans) generating billions of records of data. Companies can make use of this data to gain more profits (ROI). We can’t understand any thing from a million record data set. We can only understand after summarizing and studying the patterns in the data. Many organizations use their own data (Sales data in Retail, Transaction data in Banking). And derive the insights to increase their revenue, efficiency, or to do some promotions.
Approach to do Data Analysis
Generally we follow the below approach while analyzing the data:
Understanding the Problem and Data: It is important to understand the business questions or requirement. And also we need to understand the data. And we find important variable to use in our analysis.
Data Collection: If you have historical data, then we can go for the next step. In other cases we collect the data before proceeding for the next step. For example, If you want to analyse how a particular Super market is performing from the last 2 two years. We can study the historical data. i.e; sales data to draw the conclusions. In the second case, If you want to study how your customers are satisfying with your service. You have to collect the data (By asking the questions face to face, or by launching the surveys) to analyse the data.
Cleansing and Formatting the Data:
Once our data is ready, the next step is cleaning and formatting the data. We can not use the raw data which we have received as a input data. We have to study the data to find if there are any missing or wrong data points. And we also format the data in the required format for Analysis. We follow many approaches to clean the data. For examples, we can generate simple frequency tables and see how the data is. Or we can plot the charts (generally scatter chart) to see if there are any outliers.
Tabulation and Statistical Modelling: Once we are completed with the data cleansing, we go for tabulating the variables. We can study the data to draw the basic observations. And based on the requirement we can apply statistical techniques to understand and interpret the data. We will see these techniques later in-detail.
Interpreting or Recommendations: Based on the outputs generated in the above step, we will analyse the data.And we will write our recommendations (generally it is in Presentation or Dashboard) by following the assumptions. And we send it to the executives to take the decisions to solve the business problem.
Different kinds of Data Analysis – By Request/ Reports
- Trackers: These are the standard reporting requests to see how the different metrics are performing in a particular time period. Generally trackers are generated Weekly, Monthly, Quarterly and Yearly basis.
- Ad-hoc Requests:These requests help to quickly understand particular patterns in the data. These are mostly based on a category. These reports will be straight answers for the Ad-hoc requests.
- Drill-down Analysis: It helps to deeply understanding the data to answer a particular problem. These are summary reports based on the combinations of the variables or they can be a simulation tools.
- STATISTICAL Modeling: This will helps understanding the customer groups and their behaviors. Generally we use large amount of the data and draw the conclusions by applying the statistical techniques.
Different kinds of Analytics – By Domain
We can classify analytics as follwing types based on the domain it is applied. We will discuss more on this in later topics…
- Marketing Analytics.
- Customer Analytics.
- Risk Analytics.
- Web Analytics.
- Human Resource Analytics.
- Fraud Analytics.
- HealthCare Analytics.
- Financial Modelling.