Essential types of data analysis
Here’s a list of definitions for the most important types of data analysis:
- Exploratory – ‘How should I use the data?’
Exploratory data analysis (EDA) techniques are used by Data Analysts to investigate data sets and summarise their main characteristics, often employing data visualisation methods.
This helps determine how best to manipulate data sources to get the answers you need, making it easier for Data Analysts to discover patterns, spot anomalies, test a hypothesis, or check any underlying assumptions.
- Descriptive – ‘What happened?’
Descriptive analytics is a simple, surface-level form of data analysis that clarifies what has happened in the past. This involves using data aggregation and data mining techniques.
For instance, a company that monitors its website traffic might mine that data and find a day when the number of visitors dipped dramatically.
- Diagnostic – ‘Why did it happen?’
Once an anomaly has been identified, a Data Analyst will then look at additional data sources which might tell them why this occurred. The analyst is searching for causal relationships within the data, which could mean using probability theory, regression analysis, filtering, or time series analytics.
Following our example, the Data Analyst might consult data about the company’s day-by-day advertising spend and discover that certain advertising channels were switched off on the day the website traffic decreased.
- Predictive – ‘What is likely to happen?’
This is when Data Analysts start to come up with data-driven insights that a company can act on. Predictive analytics estimates the likelihood of a future outcome based on historical data and probability theory.
While predictive analytics can never be completely accurate, it does eliminate the guesswork from making crucial business decisions.
Using the example above, the Data Analyst could make a reasonable prediction that temporary reductions in advertising spend are likely to yield a short-term drop in website traffic.
- Prescriptive – ‘What’s the best course of action?’
Prescriptive analytics advises a business on which course of action to take and aims to take advantage of any predicted outcomes.
When conducting prescriptive analysis, Data Analysts will consider a range of possible scenarios and assess the consequences of different decisions and actions. As one of the more complex forms of analysis, this may involve working with algorithms and machine learning.
Using our example, the Data Analyst might recommend that the business maintains a more even day-by-day advertising spend in order to generate consistent levels of website traffic.
- Inferential – ‘What are the larger implications?’
When conducting people-focused data analysis, you can normally only acquire data from a sample group, because it’s too difficult or expensive to collect data from the whole population you’re interested in.
While descriptive statistics can only summarise a sample’s characteristics, inferential statistics use your sample to make reasonable guesses about the larger population. Though data might have been collected from a hundred people, you could use inferential statistics to make predictions about millions of people.
With inferential statistics, it’s important to use random and unbiased sampling methods. If your sample isn’t representative of your population, then you can’t draw large-scale conclusions.