Last week I was quite busy with MongoDB course. And even though I have plenty work with final exam which contains 10 separate works I am little bit fed up of databases at all. No matter if NoSQL or SQL which I work with really intensively in last week before Christmas.
I moved my sight to the programming language and was looking for some practical hands on experience. I have been lucky and found Data Mining: Discovering and Visualizing Patterns with Python, when I was looking for new articles and resources about data science. When I found this I was so excited! It's like kill many data science skills with one stone. I can play with Python instead of the database, work with new data and visualize it. Good, let's see how it works.
I have installed Python 2.7, Eclipse IDE Kepler and PyDev plugin. What I needed to install were numpy and pylab (resp. matplotlib and dependent packages...).
Referential Card is done as step by step commented tutorial with code snippets. Source data are about three species of Iris flowers with 50 records. Which is enough and you will use it for learning how to:
- Import data (web source)
- Visualize them
- Classify and cluster them
- Discover relationships (regression and correlation measures)
- Reduce the dimensionality of the data
- Analyze structured data (Mining networks)
You don't need to know Python language in deep detail, every code snippet is functional, so it is just enough to copy and paste it and run it. What you only need is installed Python and required libraries (which on Linux or Mac is easy and in case of Windows helps e. g. Unofficial Windows Binaries for Python Extension Packages).
And of course, what you really need is to thing about what your are doing, be precise in reading comments around code snippets and in case you want to learn something more experiment with code setup or even take another data and experiment with it.
Žádné komentáře:
Okomentovat