SQL or MongoDB? How Social Marketing companies choose

While technology like http://www.aliyun.com/zixun/aggregation/13461.html ">mongodb,mapreduce" is seen as a hallmark of innovation, companies rely on sophisticated technologies such as SQL. Mining important business models from these data becomes very difficult when the enterprise needs to collect more data formats, especially when new formats are not supported by SQL.

When faced with such a problem, marketing analyst SumAll chose a web-based Data warehouse to provide SQL query support for non-SQL data. It is understood that the company has chosen a Bityota solution that enables fast interactive analysis of JSON data stored in MongoDB.

SumAll provides users with a mining service for social media and Web application metrics data, and according to CIO Korey Lee, companies want to use SQL to generate reports that help customers better identify social trends and measure their performance with peers in social media. They support mainstream web apps, including Twitter,instagram and Facebook.

Bityota's software services can build a mapping layer on top of MongoDB that can query and aggregate data, Lee said. SumAll tracked billions of of social media data, and Lee thought MongoDB's scalability was excellent, but there were some difficulties in analyzing and aggregating MongoDB data.

"MongoDB has its own mapreduce function," Lee says: "Our programming environment often uses Hadoop or MongoDB to handle a huge dataset, but mapreduce with MongoDB, at least for us, has some difficulty expanding." ”

SumAll generates an analysis dashboard that allows customers to evaluate the outcome of marketing projects and ROI across multiple social media platforms. Help customers understand which platform to promote better, for example, for some companies, Instagram effect is stronger than Facebook.

JSON and NoSQL

The need to analyze unstructured data from MongoDB or Hadoop data sources is growing, BITYOTA CEO Dev Patel says he wants to set up the company to meet the need to integrate different types of data, A cloud-based Data Warehouse can help customers build such applications.

The NoSQL database supports the fact that the data schema is created, allowing users to open the system and accept all types of data. SumAll and many similar companies want to be able to query these web-generated nosql data through sophisticated SQL. "The NoSQL database is popular because companies need more flexibility, especially for mobile or web-side applications," says Patel. However, this does not mean that traditional database technology will die out, and we see that there are many areas where the need for SQL is very strong. ”

In the modern NoSQL data collection, JSON-type data occurs most frequently. With the advent of JavaScript development, JSON has become the best alternative to XML.

At the same time, the MongoDB database is also starting to pop up, which is often used to store JSON-formatted documents. And many traditional database vendors are not ignoring this trend, including IBM, Informix, Microsoft and Oracle and other companies are beginning to support the processing of JSON data in different ways. There are a number of customers like Sunall who want to apply traditional SQL to NoSQL data.

Lee said SumAll's choice of Bityota data warehousing services was also considered, and it was really the best choice to handle new data in a familiar way.

It is understood that SumAll's MongoDB chose Rackspace's version of Objectrocket, which mainly provides MongoDB as service (MongoDB as a services). According to Lee, speed is the main indicator of sunall selection, and now the company can make an on-demand query of the complete dataset at any time.