Data Analysis Explained, from Types to Tools
Data analysis is simpler than you think. Find out what data analysis is, how it’s performed and which data analysis tools are most commonly used.
What’s one thing hospitals, universities, online retailers and government agencies have in common? They all rely on data analysis to learn from past events, perform their best in the present and achieve their goals in the future.
That’s because data analysts don’t just look at data. With the help of various tools, they also prepare, clean, process and interpret it so the organizations they’re working with can make smarter decisions and forge a more successful path.
Quick Links
What Is Data Analysis?
You’ve probably already guessed that data analysis involves, well, analyzing data. But what does it actually entail?
While the exact tasks involved in data analysis vary based on the task at hand, it can involve inspecting, processing, cleansing, condensing, interpreting and extracting useful information from raw data.
As the U.S. Office of Research Integrity (ORI) explains, data analysts are able to accomplish those goals by applying logical and statistical techniques.
Data Analysis in Action
Need an example of data analysis? No problem.
Let’s say a hospital wants to know how many staff members it should have working on any given day. Too many, and the hospital will lose money. Too few, and patients won’t receive the care they need.
To learn the answer, a data analyst could collect as much admissions data as possible. Then, they could evaluate the patient-to-staff ratio at different times of day, days of the week and so on. Based on their findings, they could help predict when the hospital should have more staff members on-site and when they should bring in fewer people.
That’s not just a hypothetical example, either—data analysts did exactly that for the largest university hospital in Europe, Assistance Publique-Hôpitaux de Paris (AP-HP).
Data analysis isn’t just applicable to healthcare, though. It’s useful in just about every industry imaginable, from manufacturing to retail to transportation. Knowing that, it’s no wonder that overall employment of statisticians (a category to which data analysts belong) is projected to grow by a whopping 33 percent from 2020 to 2030.
Types of Data Analysis
It’s important to know that not all data analysis is the same. In fact, there are several main types of analysis, all of which involve different sources and processes.
Descriptive Data Analysis
If a data analyst is asked to perform descriptive analysis, that means they’re tasked with describing and summarizing raw, quantitative (i.e., numerical) data.
For instance, a data analyst could use descriptive analysis to summarize the average age, income and gender of a business’ customers.
Exploratory Data Analysis
Similarly to descriptive data analysis, exploratory data analysis summarizes quantitative data. The major difference is that it does so via data visualizations rather than statistics.
In other words, a data analyst who’s performing exploratory analysis would turn their findings into easy-to-understand charts and graphs instead of long spreadsheets filled with numbers.
Secondary Data Analysis
When a data analyst uses data gathered by someone else to perform their analysis, that’s known as secondary data analysis.
So a data analyst asked to analyze old data gathered by other analysts would be engaging in secondary data analysis, while one who’s gathering previously untouched data would be performing primary data analysis.
Diagnostic Data Analysis
If a data analyst wants to know why something happened, then they use diagnostic data analysis to find out.
For example, an analyst could use diagnostic analysis to determine why a company’s sales increase during certain months for no discernible reason.
Prescriptive Data Analysis
Sometimes data analysts use the insights they’ve gleaned to make recommendations for the next course of action. When they do, it’s called prescriptive analysis.
For instance, a data analyst might analyze a school’s attendance data and use their findings to recommend that class sizes be kept as small as possible in order to increase attendance.
Predictive Data Analysis
If a data analyst wants to use data to make predictions about the future, then they’ll use predictive data analysis.
An analyst could use qualitative (i.e., descriptive) data, such as customer feedback surveys, to predict which products will sell the best in the coming year—that would be considered predictive data analysis.
Big Data Analysis
As its name suggests, big data analysis involves analyzing very large data sets, whether to gather basic statistical information, make recommendations, predict future trends or any other purpose.
A data analyst who analyzes data gathered from the patients of a nationwide network of hospitals is performing big data analysis, as is one who analyzes traffic patterns in a metropolitan area.
Data Science Without Coding
The No-Code Data Science Masterclass for Business Analysts & Executives
Tools of the Trade
Whether they’re working from home or a corporate office, data analysts can use a wide range of tools to perform their analyses accurately and efficiently.
Microsoft Excel
For data analysts, the spreadsheet program Microsoft Excel is an indispensable tool. It’s usually best for small to medium-sized data sets, and one of its biggest benefits is that you don’t need to know how to code in order to learn how to use Excel like a pro.
Python
The general-purpose programming language Python is often used by data analysts, sometimes in tandem with add-on tools like pandas which are designed to make data analysis easier.
R
Another programming language often used for data analysis is R, which was created specifically for statistical analysis, graphics and reporting.
Structured Query Language (SQL)
Designed to manage data held within relational databases (i.e., collections of data that are related to one another), Structured Query Language (SQL) first appeared in 1974 and is one of the oldest programming languages still in use today.
Apache Spark
Rather than being a programming language itself, Apache Spark is an engine that accepts multiple languages and can be used for data analysis, engineering and science, as well as machine learning.
Microsoft Power BI
For data analysts looking to transform data into aesthetically pleasing charts, graphs and diagrams, Microsoft’s data visualization tool Power BI is invaluable.
The Data Analysis Process
So how do data analysts actually go about doing their job?
Depending on the type of analysis they’re going to perform, the process usually involves some or all of the following steps.
Gathering Requirements
Before digging into even a shred of data, every data analyst must first understand why they’re analyzing the data to begin with. They should also note the desired end goal of their analysis, whether it be polished visualizations, custom-tailored recommendations or detailed predictions.
Gathering Data
Once an analyst has nailed down their requirements, they can start collecting the data itself. Depending on the task at hand, this can involve a wide array of sources, from social media insights to customer surveys to paper records. The analyst will typically import all the data into their preferred program, such as Microsoft Excel.
Cleaning Up
No dataset is perfect right off the bat – they all include imperfections such as blank spaces, duplicate entries and the like. It’s up to data analysts to remove those imperfections in the interest of avoiding complications down the road.
Performing Analysis
After thoroughly cleaning their data, data analysts are finally able to start analyzing it. They can do so using any of the tools outlined above, whether it’s a language like Python or an engine like Apache Spark.
Interpreting Results
Now it’s time for what’s arguably every data analyst’s most difficult task: interpreting the results of their analysis. This requires plenty of careful consideration, and may involve creating visualizations, identifying trends or forming recommendations.
Want to Learn Data Analysis? You’re Already on Your Way
If you’re thinking about becoming a data analyst, it may seem like a daunting task. But the truth is that you don’t have to be a master programmer or expert mathematician to do so – with the right data analysis tools and some helpful instruction, anyone can learn data analysis.
So whether you’re looking to start a new career or learn a new skill, you’ve already taken the first step of becoming a data analyst by learning what data analysis is and how it’s done.
Discover Online Classes in Data Science
Machine learning, data visualization, programming, and more.
Carrie Buchholz
Carrie Buchholz is a freelance writer who lives in Northern Colorado with her husband and dog.
Try Skillshare for free! Sign up for a 7 day free trial today!
Get Started- Unlimited access to every class
- Supportive online creative community
- Learn offline with Skillshare's app
