Over the last decades statistical analysis becomes more and more complex. Statisticians needs for software and machines to compute large matrices with complex operations arises. Since measurement has become more accurate, storage is easy to access/ share and the awareness that tracked data can be useful later for researcher of all fields, the amount of data available is tremendous. Therefore, modern statistical software also needs handle with large dataset, while providing efficient algorithms
Statistics Software available:
- Proprietary
- SPSS
- STATA
- GRETL
- EViews
- GAUSS
- Matlab
- ...
- Open Source
- R
- Python
- Julia
- Octave
- ...
Currently, *Open Source* software is very popular:
- free to use
- support in forums from statisticians, scientists and professionals
- infinitely expandable by user-based packages
- very easy to start group project since it's free
R is widely used in all fields, providing a large amount on packages and with a very responsive fan-base providing support. It's still very slow in computing, requires a lot of computing performance and is not as easy to read the code as Python. For example, R-code requires many different *brackets* depending on their usage. Mixing up these *brackets* or missing one leads to errors in the code.
Here Python shines with code easy to read and as a programming language mainly used to engineer software it provides a huge potential. Here the first difference appears. Since, Python is a programming language we first need to load packages which include functions and operations we later want to use for our analysis. For users of proprietary statistic software it might be uncontentious, to load packages to compute basic statistics.
We use Python together with the Jupyter-Notebook Editor. Further information regarding the Editor can be found [here](https://www.dataquest.io/blog/jupyter-notebook-tips-tricks-shortcuts/)