Hi Aaron,

There is not a big emphasis on setting up the database itself, since we mostly assume it is set up.  Maybe not a good assumption.

There are example connection strings in the notebooks themselves.

We provide a Docker image with necessary dependencies required to compile and test MADlib on PostgreSQL 9.6. 
https://cwiki.apache.org/confluence/display/MADLIB/Quick+Start+Guide+for+Developers#QuickStartGuideforDevelopers-Dock

https://cwiki.apache.org/confluence/display/MADLIB/Installation+Guide
https://cwiki.apache.org/confluence/display/MADLIB/Quick+Start+Guide+for+Users
https://cwiki.apache.org/confluence/display/MADLIB/Quick+Start+Guide+for+Developers

If you would like to add more content, or modify what is there, to make it more accessible, that would be great.

Regarding the forum of deeplearning.ai courses, do you have a link?  This link leads to the signup.

Frank




On Wed, Aug 30, 2017 at 11:57 AM, FENG, Xixuan (Aaron) <xixuan.feng@gmail.com> wrote:
This is great!

Do we have documentation about setting up databases to connect with Jupyter Notebook to run? I think not many python users have done that.

The MLP examples are especially interesting. I suggest we share them on the forum of deeplearning.ai courses. There are so many enthusiasts who may be interested in these.

Aaron

On Wed, Aug 30, 2017 at 11:39 AM, Frank McQuillan <fmcquillan@pivotal.io> wrote:
Hello,

Now that 1.12 has shipped, I want to remind you of Jupiter notebooks with the new 1.12 features posted to:

The goal of these data science notebooks is to help you get started on the new features by showing examples of usage that you can copy.  Many of them reflect the examples in the user docs at:

The new 1.12 notebooks that have been added are:

neural nets (general MLP usage)

neural nets (demo uses the popular MNIST dataset, which consists of 70,000 hand written digits and is used for classification)

graph/all pairs shortest path

graph measures (closeness, diameter, average path length, in-out degree)

graph/breadth-first search

graph/weakly connected components

stratified sampling

train-test split

If you have your own examples that you would like to add the repo, you are most welcome to do so.

Frank