Drupal Data Mining for HolaMun2.com

The Need for Data Mining

Mun2, a part of Telemundo cable, is a channel created for America’s Latino youth. The corresponding Drupal website serves the same demographic (American Latinos ages 13-18) and provides information on the tv lineup as well as news and related interactive content.

In an effort to anticipate and respond to user trends, and in order to draw and retain visitors, HolaMun2 project managers needed a robust business intelligence platform to query visitor data and perform analysis using visualization tools.

Data Mining in Drupal

However useful visitor data may be, significant challenges face any organization wishing to extract it. In some cases, a robust data mining solution requires a proprietary system with its own API, hardware requirements, license fees, and ongoing maintenance.

Drupal alleviates some of this pain because it offers organizations the option of building integrated data mining platforms within an open-source framework. In the case of HolaMun2, Achieve Internet built a web-accessible data-mining package that allowed managers to review visitor data through Drupal. The solution visually renders the data to speed up comprehension and make decision-making easier. The Achieve team satisfied the immediate reporting requirements and also created an extensible system that enables low-maintenance query-building for changing business needs.

Business Intelligence Architecture

Achieve and HolaMun2’s internal web team agreed that performing data mining queries on the site’s Drupal database would be ill-advised. First, the queries would be complicated because the Drupal database is optimized for content management, not business intelligence. Second, performing queries on the live, user-facing site would negatively affect performance.

Accordingly, the Achieve team built a custom database designed in a more mining-friendly format. This custom database pulls data from Drupal and re-deposits it into tables optimized for mining with page_view keys. This information is then accessed via a separate Drupal site and displayed in a user-friendly fashion.

Importing Site Data

Using the existing query requirements, and a list of future enhancements, the Achieve team designed the custom database and mapped the Drupal fields to their new homes. Achieve also created custom Drupal modules to handle the transfer. When import is clicked, the modules connect to the site database, present the user credentials, extract the data, and insert it into the custom database.

Though this data import could be scheduled in cron, HolaMun2 site admins are comfortable, for now, with performing this function manually as required. The UI for this function is straightforward, simply requiring the user’s credentials and location of the database.

The interface pictured is accessible on the Data-Mining Drupal site. Note that the user can specify the database name, host, and port. Although the current setup serves just one site, the Achieve team added this flexibility, and the mining platform can house data from multiple sites in the future, re-using the importing process.