Skip to content
Kezdőlap » What is data analysis? 

What is data analysis? 

One purpose of data analysis is to extract information from the available data and use it to speed up decision-making. A recent survey by the International Data Corporation (IDC) showed that companies using analytics tools perform on average 2.5 times better than laggards on 6 of the 12 surveyed key business metrics. The study also claimed that companies are increasingly willing to spend resources to improve analytical capabilities. 

Data analytics can help us describe the past, predict the future and make recommendations to optimise result. 

Descriptive analysis 

Descriptive analysis helps us to understand the past. We collect historical data and use them to answer the questions “what” and “how much”, like “how have quarterly revenues evolved over the past two years?”. 

Diagnostic analysis 

In diagnostic analysis, we are looking for answers to the “why” of the information extracted with the help of descriptive analysis. We try to find the patterns and the relationships between data. “How have the changes of the exchange rate affected quarterly revenue trends?” 

Predictive analysis 

With the help of predictive analysis, we use the data of the past to predict the future. “What is the expected revenue for the next quarter?”  

Prospective analysis 

Prospective analysis is used to make recommendations for future decisions. “If we increase advertising budget by 10%, the quarterly revenue will increase by 5%.” 

Important business data models 

  • Regression 

How are the variables related, how a given independent variable affects a dependent variable? “How do revenues (dependent variable) change as advertising spending (independent variable) increases?” Using this method, we can provide easily interpretable information for decision making. 

  • Monte Carlo simulations 

Monte Carlo simulations can be used to estimate the uncertain outcomes of an event. This method often used in risk analysis. 

  • Factor analysis, principal component analysis 

These methods aim to reduce the dimensions of the data. Factor analysis is used to find the underlying structure and relationships of variables. In principal component analysis, we look for the most important components and variables. 

  • Cohort analysis 

In cohort analysis, we do not look at the data as a whole, but divide it over time and perform the same analysis on these cohorts. It allows us to examine how the behaviour of a group of users changes over time. 

  • Cluster analysis 

In cluster analysis, we group our observation units into groups so that the group members are as similar as possible and the different groups are as different as possible. For example, you can use this method to find customer groups. 

  • Time series analysis 

In time series analysis, we look at the change in our data over time, looking for trends in our data. 

  • Emotional analysis 

By analysing the emotional content of different texts, we can find out how reviewers feel about our products and service. 

Some key points for a successful data project 

  • Coordination 

Before introducing new technologies, focus on better coordination of people and processes. Strive to ensure that data is central to business objectives and that it is easily accessible. 

  • Clearly defined goals 

If our objectives are not clearly defined, it is easy to work in vain, so it is important to make clarify our purpose and the usage of the results. 

  • Identifying critical points 

Do we have the right quality and quantity of data for the purpose of the project? Do we have the right methodological knowledge? 

  • Scalability 

When choosing tools, make sure that it can manage increasing data volumes, deeper analyses and growing numbers of users. 

  • Compliance 

Pay attention to what data you use, what industry-specific requirements you need to comply with, and what laws and regulations apply to you as you manage your business data. 

  • Refining models 

As time goes on, don’t be afraid to change previous models if you feel the need. 

  • Standardized reporting 

Strive to produce similar reports and visualisations for all users, to minimise potential misinterpretation of different formats. 

  • Involve colleagues 

Make the importance of the data visible and inform colleagues from other departments as well. 

Source: https://venturebeat.com/data-infrastructure/what-is-data-analytics-definition-models-life-cycle-and-application-best-practices/