Today’s entrance is to introduce you to a new web I met a few days ago, and it aims to be “YouTube of data”. The website is quandl.com.
Quandl.com is a website where you can find thousands of datasets from several companies and organisations, ready to use, and free. You can upload your own datasets. Until here, everything seems really cool, But, there is even more! The feature I like the most is to register using your GitHub accounr, but also you can do it with your LinkedIn or Google account.
Once logged in, there is a wonderful help in five easy steps to show you how to work inside Quandl. I must admit I like the desing, and the easy of use!
Working with datasets
One of the main feature of Quandl’s datasets is the representation the data, giving the user the option to dowlonad the data is several formats (one really cool feature!). A extra feature I also liked is that a user can send a message to the owner of the data in case of errors, corrections, …
If you are interested in financial data, one of the companies more active is NASDAQ, that is collaborating with this website.
Using the API
The most interesting feature for developers os the API included in the wabsite, making easy for devs to use datasets. Here you can see how easy is to connect using Python: http://www.quandl.com/help/python. The API is also available in several languages, like R, C#, MatLab, … You can see the complete list of languages here.
But, if you like to use Excel, you have a really cool extension available for you, and that you can find all documentation here.
So, …. now you have the data. It’s your turn!!
If you need a good example, here you have a Python Notebook (http://nbviewer.ipython.org/github/jrosen3/cs109annstock/blob/master/Master.ipynb) where the objective is a neural network app, predicting the stock market, using the Quandl data!
PD: By the way, while preparing this entrance I found a bug I already reported, and I hope to be fixed soon: http://www.quandl.com/search/españa