Ideas for python.ipynb

[NikeCW]

issue for Statistical_Analysis_Summerschool_22_python

in general: GUI would be nice

in 2.1 Package installation

• comments are written for the google colab workflow, maybe rewrite them for users of the jupyter notebook

• # importing necessary modules --> "import skbio # Don't import on Windows!! --> write if statement for this instead of using the "import" command

• Question: What should Windows users do at this point? skbio is needed at a later point in the script (in 3. Data-cleanup)

in 2.2 Setting a local working directory

• #enter the directory for the data files: Was not working for me, changed it to:
  data_dir = r"folderpath"
  os.chdir(data_dir)
  (deleted "input" and added raw string)

• same here #enter the directory for the results: changed it to:
  result_dir = r"folderpath"
  os.chdir(result_dir)

in 2.3.1 Loading the data: Use one of the methods

• " # optional analog match file an_file = filenames[3]" --> file is not optional, could not run program without it
• maybe there is a better way than inserting the number of the filenames in the list manually
• add comment: numbers of files in the list start with "0" (0 is first file, 1 is second file,...)

[Schampo]

To this point: Question: What should Windows users do at this point? skbio is needed at a later point in the script (in 3. Data-cleanup)

I have this same issue with 'import skbio' when running the code in google colab and then when I ignore it and proceed, my visualization for the PCoA comes up empty. I assume many of the target audience will be windows users like me so a work around will be great. My data frame looks fine after merging quant_table and metadata.

[amcaraballor]

Agree with Schampo's comment. Specifically, in the 4. Multivariate Analysis, Step 34: Conduct PERMANOVA Test, requires skbio package ('import scipy.spatial.distance as distance', and 'import skbio.stats.distance'), however as found in http://scikit-bio.org/, Note: scikit-bio is no longer compatible with Python 2. scikit-bio is compatible with Python 3.8 and later. Definitely, an alternative for users needs to be implemented. Overall, great job.