Аннотация

I HAVE BEEN USING PYTHON IS SOME PROFESSIONAL CAPACITY SINCE THE TURN OF the century. One of the trends that I have seen in that time is the uptake of Python for various aspects of "data science"- gathering data, cleaning data, analysis, machine learning, and visualization. The pandas library has seen much uptake in this area.
pandas 1 is a data analysis library for Python that has exploded in popularity over the past years. The website describes it thusly:
“pandas is an open source, BSD-licensed library providing high-performance, easy-to-use data structures and data analysis tools for the Python programming language.”
-pandas.pydata.org My description of pandas is: pandas is an in memory nosql database, that has sql-like constructs, basic statistical and analytic support, as well as graphing capability. Because it is built on top of Cython, it has less memory overhead and runs quicker. Many people are using pandas to replace Excel, perform ETL, process tabular data, load CSV or JSON files, and more. Though it grew out of the financial sector (for analysis of time series data), it is now a general purpose data manipulation library.
Because pandas has some lineage back to NumPy, it adopts some NumPy'isms that normal Python programmers may not be aware of or familiar with. Certainly, one could go out and use Cython to perform fast typed data analysis with a Python-like dialect, but with pandas, you don't need to. This work is done for you. If you are using pandas and the vectorized operations, you are getting close to C level speeds, but writing Python.