All geek questions in one place

What are some good statistical visualization structures / libraries for user data?

We collect user data on our site for a while, and now we want to provide data in an accessible way.

We have a database full of data, we are just looking for a framework that we can dump into all our data and be able to visualize it.

Some requirements:

  • It should be able to sort and filter by several parameters (for example: by user, by type of action, by date, by metadata).
  • Must be able to export data representations to csv / xml files.
  • It must be presented via the web interface (without desktop applications).
  • Must be used with Scala, Java, Python or Ruby.

We looked at using Vaadin and we enjoy the amount of control we can have, but I would like to explore other potential solutions.

Are there any suggestions for frameworks that can help us display our statistics?

+4
source share
3 answers

I have a lot of fun working with google visualization and the java library provided by Google. I even wrote an adapter for it - in Scala - to work with LucidDb, because by default the DB adapter is only available for MySQL.

The traffic schedule, in particular, is very beautiful.

Update

Lucid is a great way to handle massive amounts of data while working happily with sizes larger than 1Tb. It is column oriented (e.g. vertica or SybaseIQ), so you won’t incur any I / O costs for collecting entire rows of data to select a couple of columns.

It is also useful that Lucid is queried via SQL and has a jdbc client driver combined with google's Java visualization library, this is a perfect match, and the library can translate most of its SQL query format, thereby minimizing the amount of processing that it should perform after request.

I run this inside the Scalate framework, creating all the javascript to dynamically display the charts. Thus, I can easily change the query parameters based on other controls on the page and display the new data in the chart without having to reload the page. Depending on how heavy your site will be, Lift may be better for you.

It took me about a week to set up, although I had previous experience with the google library.

+4
source

I think Pentaho BI should fit your needs. The suite is implemented in Java! Data Integration pentaho (kettle) should be able to handle all data formats (sql-db, xml, csv, txt, etc.).

There is also a Pentaho Community Edition (open source and free) . its worth a try!

+1
source

Infobright is also a popular choice for storing huge amounts of data and pulling it for analytics. This column is oriented and uses a grid data architecture for quick and easy query resolution. It is available both with open source and corporate versions.

Infobright offers a virtual machine with Pentaho, Jaspersoft, and BIRT at infobright.org. If you want to play with these BI tools, I recommend you get started.

+1
source

Source: https://habr.com/ru/post/1342048/

More articles:


All Articles