Whether or not you are used to working with data on a day-to-day basis, you may have found it difficult to understand its meaning. I think we've all suffered the endless data sheets where it's impossible to know if a number is a good thing, a bad thing, or simply has a value that leads to concrete action.
According to Wikipedia's definition of Data Visualization, it "refers to the techniques used to communicate data or information by encoding it as visual objects (e.g., points, lines or bars) contained in graphics".
If we use an everyday language, we can say that Data Visualization is the act of communicating data in a visual way so that they are understandable, easily interpreted and can be analyzed visually, to discover patterns like:
The first step to build a Data Visualization with any tool is precisely the data. Today there are literally hundreds of different types of data sources, which are used more or less often depending on the type of company and what data you want to analyze.
These are some types of data sources that a company can use:
Databases, containing different types of information depending on whether it is an ERP (Enterprise Resource System), a CRM (Customer Relationship Management) or any other type of application (support, document management, payment management, etc.). They are usually the main data sources used by companies and in which they have updated information every day.
Files of various formats: CSV, Excel, JSON, XML, etc. These types of data sources are very common, as it is easy to export our data in any of these formats from any tool.
Cloud apps. There are hundreds of cloud applications that allow access to data that is managed in various ways. Examples: a cloud billing application, or a social network you use in your company.
APIs. Today's world is (and will be even more) hyperconnected. More and more services, apps and transactions allow access to data through various types of APIs. The most common are REST APIs, but lately big services (and also Biuwer) have bet on GraphQL APIs.
Accessing and interconnecting with the data source depends on each case, but it is an indispensable requirement. Depending on the data management approach you use in your organization, you may have an automated system that integrates all data of interest for analysis in an Data Warehouse (DWH).
From our experience and whenever possible, we recommend processing and preparing data to be stored in the corporate Data Warehouse, for many reasons, among which we highlight:
If your organization does not have a data warehouse, the data analysis tool you use has data connectors that allow direct access to raw data. It may take longer for your data queries or there may be some limitation in the calculations you want to make, but you will be able to create your Data Visualizations.
A Data Visualization is effective when it fulfills its purpose, meaning, it allows users to easily interpret the information displayed, by asking more questions about the information displayed than how it is shown (e.g. the colors chosen).
On the way from raw data to seeing it in a visualization, this must be taken into account:
As in all areas with a certain complexity, there are many types of graphics and visualizations that can be created.
Here is a list of very interesting sites that include guides to choose the type of visualization to use, from various points of view:
As you can see there are many types. Many focus mainly on graphics, but we should not forget other types of visualization. We highlight the following ten visualization types:
There are multiple uses, probably as many as needs. Each type of data visualization can be used in different ways, some of the most common are:
An online store wants to know how is evolving the average amount of the products they sell. The reason of doing this question is either to bet for massive selling of cheap products or selling expensive products in less quantity.
Starting from a database in Azure SQL Database, we follow the next steps in Biuwer:
The result of point 3 is a detailed data table, but we can’t extract any insight from here, at least at first sight.
However, after building the data visualization, it’s easy to see that the sales amount per order is decreasing and decreasing over time, with two major steps, that must match with two internal events in the company when they decided to change the product and sales strategy.
As you can read in this article, data visualization is a wide and versatil discipline and thousands of specialists are now needed in companies to bring sense to the massive amount of data generated every day.