PyData Evolution

Question

I read about PyData in several places (like here ), but I'm still not sure what the term really means.

Is PyData an official? (for example, is there a fund that owns / maintains PyData.org ?). Is this just a conference? Or is it basically a term that is used to refer to a list of Python packages?

Also, which packages are considered the main part of the PyData ecosystem? Is this some kind of package that can be used to work with data? (that would be pretty general). Some packages that I discovered are usually related to PyData:

• Numpy
• Scipy
• Pandas
• Scikit-learn
• NLTK
• Pymc
• Numba
• Blaze

Does this list correspond to the group of packages commonly associated with PyData? Or are there any important omissions?

Finally, to what extent does the PyData ecosystem support Python 3.x ? Can we assume that most of the PyData ecosystem is compatible with Python 3.x? If not, which packages do not yet support it?

Answer 1

All Google search results show that PyData is just a conference. Here about the page . This does not apply to the general community of people using Python for data analysis, or to the installed tools for this.

Answer 2

PyData is a series of on-site meetings and conferences organized with the help of NumFOCUS , a nonprofit group that supports open source scientific software. There are over 100 locally organized PyData meeting groups in the world. Major PyData conferences are also held throughout the year in Silicon Valley, Boston, New York, London and elsewhere. Many of the conference organizers are based in Austin, Texas, since NumFOCUS was originally founded at Anaconda, Inc. , a company founded by Travis Olyphant and Peter Wang. Lea Silenus is the main organizer of all the conferences, but they also hire local volunteer organizers to help with most of the logistics for each event, and I volunteer time to work on their website.

PyData also belongs to a community that focuses mainly on using Python for data analysis (more focused on business than SciPy, which is organized by Enthought and more focused on academic applications). There are many similarities between the two communities, however you will find other finance related topics in PyData.

PyData also applies to the packages you have listed. In addition, many people in the community use iPython notebooks to demonstrate the use of these packages. Anaconda, Inc has the Anaconda distribution available for download, which is an all-in-one installation that includes 100+ of the most popular Python packages for science, mathematics, engineering, data analysis. Anaconda supports Python 2.7 and 3.5+.

Answer 3

Although PyData itself is a forum for anyone interested in data analysis (managed by NumFOCUS), it seems that there is a concept of a “stack” of PyData, which refers to the stack of tools and libraries selected by the sponsoring company (I'm not sure what the selection criteria are) which is available for download ( https://pydata.org/downloads.html ). This is not an official list, as far as I can tell. Anaconda also has what its creators define as the PyData stack, and they are accessible through Anaconda. They are different from what is on the PyData website. As a result, it looks like an informal list with different definitions. But there are some packages like the ones you have listed that are common.