When engaging in data analysis, you should know the basic terms. Here are some that you’ll often hear:
Categorical Data: This is data that does not have a numerical meaning. Rather, it has a textual meaning like a description of a group (race and gender). Although, you can assign numbers to each of the elements.
Data Type: This is the kind of information a variable represents, such as a Boolean, integer, string, or floating point number.
Descriptive Analytics: This is analyzing data to get a better understanding of the current status of a business. Some examples of this include measuring what products are selling better or determining risks in customer support. There are many traditional software tools for descriptive analytics, such as BI applications.
Diagnostic Analytics: This is querying data to see why something has happened. This type of analytics uses techniques like data mining, decision trees, and correlations.
ETL (Extraction, Transformation, and Load): This is a form of data integration and is typically used in a data warehouse.
Feature: This is a column of data.
Instance: This is a row of data.
Metadata: This is data about data—that is, descriptions. For example, a music file can have metadata like the size, length, date of upload, comments, genre, artist, and so on. This type of data can wind up being quite useful for an AI project.
Numerical Data: This is any data that can be represented by a number. But numerical data can have two forms. There is discrete data, which is an integer—that is, a number without a decimal point. Then there is continuous data that has a flow, say temperature or time.
OLAP (Online Analytical Processing): This is technology that allows you to analyze information from various databases.
Ordinal Data: This is a mix of numerical and categorical data. A common example of this is the five-star rating on Amazon.com. It has both a star and a number associated with it.
Predictive Analytics: This involves using data to make forecasts. The models for this are usually sophisticated and rely on AI approaches like machine learning. To be effective, it is important to update the underlying model with new data. Some of the tools for predictive analytics include machine learning approaches like regressions.
Prescriptive Analytics: This is about leveraging Big Data to make better decisions. This is not only focued on predicting outcomes—but understanding the rationales. And this is where AI plays a big part.
Scalar Variables: These are variables that hold single values like name or credit card number.
Transactional Data: This is data that is recorded on financial, business, and logistical actions. Examples include payments, invoices, and insurance claims.
Conclusion
Being successful with AI means having a data-driven culture. This is what has been critical for companies like Amazon.com, Google, and Facebook. When making decisions, they look to the data first. There should also be wide availability of data across the organization.
Without this approach, success with AI will be fleeting, regardless of your planning. Perhaps this helps explain that—according to a study from NewVantage Partners—about 77% of respondents say that “business adoption” of Big Data and AI remain challenges.22
Leave a Reply