devvyn / aafc-field-data Goto Github PK

Since this software project is centred around some real world data, the example data should be from actual, scientific research projects. That way, more parties benefit from the success because the published software project and the published research project service each other's needs. Future scientific research can benefit from the leverage of the tools that have resulted from prior iterations, and therefore the benefit to scientific research is multiplied over time.

• 🐳 docker-compose.yml or script
• 💾 working directory on host
• ability to keep data in repo tracking
• 📝 instructions on starting and using JN

consider:

• 💩 users who are trapped in a crippled, Window only workplace—with Internet Explorer

As per: http://repo2docker.readthedocs.io/en/latest/usage.html#preparing-your-repository

At the same time, it makes sense to split up the README so the appropriate focus is possible in context.

This would get us:

/
|
+--binder/
|   |
|   +--README
|   |
|   +--requirements.txt
|   |
|   +--postBuild
|
+--notebook/
|   |
|   +--README
|   |
|   +--projects/
|
+--README

might be an issue with widgets in general, but ipyleaflet for sure doesn't show anything

To reproduce

Install package:

conda install -c conda-forge ipyleaflet

Example map from http://jupyter.org/widgets:

from ipyleaflet import Map

Map(center=[34.6252978589571, -77.34580993652344], zoom=10)

Expected

a graphical map

Actual

nothing, not even an error

Link from README (or more suitable document) to:

• Example of reading Excel sheet
• Using Python & pandas to Analyze an Excel File
• CAM-2017
  • instructional notebook
  • non-instructional
• combining-sweep-and-tiller

• homepage
• README

• find purest data
• align data
• plot graph

https://github.com/devvyn/aafc-field-data/projects/1

Whether by docker-compose up or docker run, the Python packages aren't installed automatically. To make matters worse, the requirements.txt isn't accessible to the Jupyter Notebook terminal.

Maybe a better Dockerfile could be added to the project, or one of the startup scripts could do the installs. MyBinder.org will use the requirements file already, so it may not be necessary to use Dockerfile.

• update README to explain how to change the requirements so that they are honoured everywhere that the notebooks in the repository are likely to be used (MyBinder.org, local container)

• replace requirements.txt with environment.yml because it works better with Jupyter
• get rid of docker-compose.yml because it doesn't work with repo2docker
• update docs
• update changelog

• identify chunk boundaries
• read data into pandas objects or standard Python data structures
• output columnar data to file that doesn't require chunked parsing
• document use of parser
  • can be in notebook or docs
• split Python tutorials apart from analysis
• update change log

It's a bit of a mess near the end. Is the appendix complete? Is there a final analysis? Should the reader be referred to the analysis notebook? (2017 CAM data from iPads.ipynb)