Best tool / programming language for data visualization

Question

So far, I have mainly used excel to build any data that I extract from MySQL using PHP (sometimes using a JS library such as HighCharts to build data). This is not an option when I need more complex diagrams / visualizations, or if there are many records (excel gets slooooow), so I'm looking for alternatives. However, there are so many options! I have downloaded and tried things like R and Processing several times, but I am not yet convinced that this is the right tool for the “simple” construction of my data in a good way.

So ... they are curious here about the best way. Learn Python? Really immersed in Java / processing? any other options? (or stick with PHP, and do you have a real flexible graphical display library ?;))

Answer 1

I think Python is great for fast and rich data processing.

Answer 2

Personally, I think that Incanter is a great set of tools that I can recommend from the bottom of my heart (I use it for visualization in my own projects).

This is a statistical computing and visualization library for Clojure - which, in turn, is a very powerful, flexible and dynamic langauge, especially suitable for interactive data experiments

Great features:

• Numbering provided parallel stallion
• Charts are created using the very complete JFreeChart library
• It also uses Processing for some visualizations.
• Clojure itself is a great language for data processing and intelligence.
• Interaction with Java is excellent, so you can easily use any Java libraries (for example, to access a database, web services, integrate message queuing, create networks, etc.).

I especially like DSL for charting, for example. to create a histogram of 1000 samples from a normal distribution, which you can simply do:

(view (histogram (sample-normal 1000))) 

Answer 3

I think it swings the clojure library for statistics and data visualization. clojure is functional and superior in data processing and analytics

Answer 4

I highly recommend using Vega-Lite & Vega to visualize data from any language you choose. Vega-Lite and Vega are based on the ideas of grammar graphics, which inspired the popular ggplot2 library. The basic idea is that data visualizations should be built in accordance with declarative descriptions of how the properties of the data correspond to the aesthetics of the dataviz. Vega-Lite and Vega, however, take it one step further by providing an interaction grammar that allows you to create interactive data visualizations and complex explorer views. Moreover, it increases the emphasis on the declarative nature of GG in the sense that the Vega-Lite and Vega specifications are described as pure data (JSON). Thus, any language that can target JSON can target Vega-Lite and Vega.

Vega-Lite is a more or less high-tech data analysis tool focused on providing high leverage and automation based on highly Spartan specifications. It compiles into Vega, which is a slightly lower level and more powerful but less automated version of Vega-Lite. Usually it’s enough to start with Vega-Lite and switch to Vega only as needed.

For more information about Vega and Vega-Lite, see: https://vega.imtqy.com .

In conclusion, I will reinforce the opinion of Meeker and Ravinder Ram that Clojure is an excellent language for data science and is constantly being improved thanks to new machine learning libraries (for example, MXNet recently received support from Clojure ). Also now there is support for most modern scientific applications for laptops (Jupyter, Nextjournal, Gorilla REPL), if this is your thing.

If you want to use Vega-Lite or Vega from Clojure or ClojureScript, you can use the small but flexible wrapper library that I wrote and called Oz:

https://github.com/metasoarous/oz

If you are interested in using Vega-Lite or Vega from other languages, there are plenty to choose from because of the ease of porting (e.g. Python, R, React).

Answer 5

Python, Octave, C / C ++, etc. For each of them there are many libraries for printing. A.

I can recommend the MathGL GPL graphics library and its UDAV interface. The latter has a different command language to build.

Answer 6

R-language is mainly used for data science - you can do complex manipulations with mathematical data and create predictive models, but this will not help you with its convenience and simple graphics. I don’t think you need Java unless you have complex business logic or multithreading.

Therefore, I would recommend sticking with Python - it is really convenient for data processing, which is easy to learn. In fact, working with data is the power of Python. You can also select several (link is a good compilation found on the Internet) data visualization libraries or try javascript from above. Do not be afraid and good luck!

Answer 7

I personally started using Python + D3.js.

Python is an excellent technology for working with data - for managing, converting and output in various formats.

And D3.js is a great library for data visualization. This is much easier than the first. Let's start with a few examples online. And most importantly, you can create interactive charts that allow you to present different views or different levels of data as you drill into it.