Data analysis to write code

This is a created article in which the information may have evolved or changed.

Recent tasks have been a bit busy, mainly in the data analysis needs of various departments, coupled with the app backend performance optimization squeeze together.

The previous data-related work was mostly passive, and each department gave the data analysis requirements, and then a person analyzed the data from ES or Hadoop, and finally drew it with tableau, then exported it to PDF via keynote and sent it back via email.

The recent adoption of a proposal is to develop an internal use of big Data open platform, enter a keyword on the Web, you can return the relevant information and analysis results. Liberate yourself from a lot of manual analysis.

Task is urgent, from zero to on-line only to one months time, and only I a developer, still can not delay normal work.

So the tradeoff is that the preliminary selection is based on the restful API provided by Tableau server, which returns the pre-generated diagram back to the Web front end. It also gets a real-time computed data from Redis, as well as information about the text classes that are captured in MySQL.

Nginx is mainly used to do security control, only allow the company intranet use, so do not need the account login module.

The overall architecture:

Today I upgraded Tableau server to the latest version and can download HD images of views based on the HTTP Restful API. Fortunately, today's upgrade, and less than one months to maintain the license expires, then will not be able to upgrade. The lower version of Tableau Server provides a weaker API and is completely unable to meet the initial requirements.

The front end provides a simple search box, and then the search for content and images will appear below, and the first version will temporarily not provide the data export function.

At present is based on Python quickly build a demo out, wait until the demo set up almost, to use Golang to do for the backstage, the main consideration is the way to pick up the lost long Golang.