BI software: Data mining now within everyone's reach
Data analysis tools are gradually invading all businesses, because although the interest is not new, software has evolved to be much more accessible. Over and above this trend, the evolution in usage is now moving towards data mining. Why and how? A tutorial on data analysis is in order: here's how.
Data Mining: definition
Data mining covers all the methods, techniques and tools used to uncover knowledge from large volumes of data. Unlike Business Intelligence, whose aim is to create reports and dashboards to understand a situation and make decisions, Data Mining aims to identify correlations and new , unknown information.
A revolution in practices thanks to online BI software
Data mining has been around for a very long time. In fact, some believe that the first data mining operations were carried out by the Chinese emperors almost 4,300 years ago. The disruptive factor that produced the latest revolution in BI is the migration of data to the Cloud.
The fact that companies' most important data is hosted in private or public Clouds is changing the game in more ways than one:
- Data can be accessed via a simple Internet connection;
- Data can be aggregated in a single location;
- Data can be read in real time;
- There is no limit to the amount of data that can be archived, either in terms of time or volume, thanks to the decreasing cost of storage.
This technological breakthrough is also accompanied by the simplification of data mining tools. The American company Zendesk, with its Bime business intelligence tool, is a perfect example. It enables web marketers, sales managers and financial controllers to create the dashboards they need to manage their business.
More concretely, this means that those concerned with the data (the business) do not need to go through a Business Intelligence Analyst to query the databases and extract the information contained in the data. Bime has over 60 connectors that allow users to 'plug in' to the data sources they need and then create their own tables or cubes in the same way as a pivot table in Excel. These data sources include
- Social networks: Twitter, Facebook, Linkedin, YouTube, Instagram, Vimeo;
- Website data: Google Analytics, MySQL;
- Static files: Excel, CSV, Google Drive, DropBox, XML;
- Marketing software: Campaign Monitor, MailChimp, Zendesk, Salesforce;
- And many other sources
From Business Intelligence to Data Mining
Using a tool like Bime not only allows you to design interactive dashboards whose data is read in real time, but also to get a foot in the door of data mining thanks to practices that are easy to put in place.
- By aggregating multiple data sources, it is possible to identify statistical correlations between remote data. For example, by mixing your Facebook data with that from your e-commerce site, you can discover that women are more inclined to make a purchase following a campaign on social networks, whereas men buy on the basis of product comparisons.
- The drill-down function breaks down the data into sub-segments. Let's take the example of a marketing campaign in France: the BI software will allow you to separate revenues by channel (Adwords, Emailing, Billboarding in public places) then take one of these channels and observe the results by city. The aim is to target marketing budgets more effectively.
- The analysis engine's algorithm also makes it possible to identify trends (stagnation, decrease, increase) as well as forecasts incorporating seasonal factors. This makes it possible to capitalise on past results to anticipate future sales, for example.
Data mining is not an easy discipline. It requires a great deal of rigour and method. On the other hand, a tool like Bime breaks down the barriers external to the user, such as accessing data without SQL queries or extraction, and manipulating data without formulas or manual calculations.
Explore your data in 4 simple steps
1. Define your problem
The first step is to focus your attention on the objective you want to achieve. This could be solving a problem (e.g. understanding why sales fall in May), optimising marketing budgets or improving a product.
It is also a question of formalising the graphs that will help to understand the problem.
2. Connecting and structuring your data sources
Software such as Bime connects to data sources in two ways: by importing a file (CSV, Excel) or with identifiers and passwords in the case of a 'Cloud' source. If you want to retrieve data from your website or any other private database, you will need to add a step to break down the fields into two groups: attributes and measures. Attributes are names, words and identifiers, while measures are numerical values.
3. Creating queries and graphing them
Creating queries and graphs is the most time-consuming stage because, as well as creating them, it calls on your analytical mind. Bime allows you to analyse a single data source or several on the same graph using Query Blender.
This tutorial shows you how to get to grips with this component:
4. Dashboards and reports
The purpose of BI software is to be able to formalise and distribute results that are rich in "actionable" information. Users add the graphs of their choice to dashboards, which can be accessed by an unlimited number of readers via logins. The great advantage of a tool like Bime over a traditional Business Intelligence solution is that it provides access to data in real time: each time the dashboard is refreshed, it is updated with 'live' data.
What are the benefits of data mining for businesses?
Data mining allows you to go further than simply understanding your business. It allows you to look into unknown areas of a market or business that could potentially revolutionise the company's future.
Given the price of a Bime licence and the time required from the user, data mining is a real opportunity to create competitiveness in companies. The collection of data and the discoveries that result from it give market share to those who take an interest in it.
Conclusion
Data mining is now accessible thanks to ergonomic online Business Intelligence tools and the prosperity of data in the Cloud. What's more, this discipline has moved from the hands of technicians to those of business managers who are more inclined to understand the information they are mining. Although the technical hurdles have fallen, data mining still requires method and rigour. Finally, data mining is a real business opportunity for companies looking to boost their growth and competitiveness.
Article translated from French