How to create visualisation on medical data?

Question

I want to create a data visualization on medical data i.e. patient's medical history, allergies one is having, any chronic patient's name must be highlighted, etc. And, there can be a separate visualisation for the medicines, which shows medicines' availability, their expiry date can be highlighted by using mouse_hover_on functionality. It will be useful for the doctors and the medical staff as well.

Problems:

• bl.ocks.org has variety of base visualization models to work upon, but the schema of database I have (given by my mentor) is a bit confusing.
• Also, he won't provide us with the actual data of our college due to its confidential nature, so can we get dummy data on the net?

Experience:

1. I have done work on Weka earlier to understand about data mining, and many different algorithms and classifiers.
2. I have built a generic data visualization model (for hierarchical data) using the D3.js
3. I have good working knowledge of Python. So, if anyone could please tell whether there is any hope for doing it(the data visualisation) in python.

Answer 1

If you have good working knowledge in Python than you are good for it. D3.js actually has a Python comparable called Bokeh. You question implies a lot of ramifications and does explain fully what you are trying to achieve so let's go by parts.

Some 2D Scientific Plotting Python Libraries:

• Matplotlib
• ggplot
• Bokeh
• Chaco
• pyQtGraph (This one has some interesting features in volume slicing if it suits you)

Some UI development Python libraries:

• PyQt5
• PyQt4 (there are significant changes to PyQt5, thus the suggestion)
• PySide (also a Qt port)
• wxPython

Some Python numerical libraries:

• Pandas
• Numpy
• Scipy
• Scikit-learn (You can do some serious data analysis with this)

Using combinations of the above you can build some very powerful software solutions. Any of the UI libraries suggested has spreadsheet widgets. There are libraries to read csv and excel, among others (JSON, HDF5, etc.).

The data problem can be resolved by either building your own randomized data or by using resources on the net like this or this.