# Analysing the Titanic accident with SAP Predictive Analysis

As part of my participation in the Data Geek Challenge 2.0 in the summer of 2013. I stumbled upon the records of the Titanic sinking from the University of Toronto. During the analysis with SAP Predictive Analysis it shows very quickly the extent of the tragedy and the low chance to survice for each passanger.

### Assigned Tags

You must be Logged on to comment or reply to a post.

Very good and clever entry.

I enjoyed the music.

Well done!

Hi Boris,

A very interesting use case. The music made me feel the Titanic was sinking once again while I was watching this video π .

Benedict

Impressive analysis and impressive musical score! Well done Boris, we will be sure to feature this entry on the Data Geek homepage, www.sap.com/datageek

Very Well done, loved the music too π  Mitesh

Blog Post Author

Thanks for the nice comments. I am glad that you like it.

greetings

boris

hi Boris,

What software did you use to record the video/audio for this presentation?

Benedict

Blog Post Author

Hi Benedict,

I used SnagIt to capture the video and Windows Movie Maker to cut it. Hope this solves your question.

greeting

boris

Thanks Boris. I feel inspired to do something in the DG challenge after seeing your video:)

Will give it a try.

Thanks,

Benedict

Nice one, it will be great if you could add some explanation on what the the tree diagram is trying to show. Me being from statistics backgound understood it, but my friend from a non statistic background did not get it so had to explain it by running an actual scenario using the data file from the source you mentioned

Blog Post Author

Hi Bimal,

as you wish a small comment about the decision tree diagramm. The decision tree is a special tool for classification in DM systems. I used the R CNR Tree Methode in the titanic szenario. The Generation is based on the top-down principle. The Starting point (the root) contains all records of the training set which is divided with the aid of the rules defined by the variables in two or more sub-nodes (sons / daughters).

As a measure for determining the best classification tree the Gini Index is used.

By analysing the solution of the diagramm you get two profiles which had a chance to survive:

Profil 1: female and part of the first class or a crew member

Profil 2: a male child and part of the second class or higher.

The quality of the solution is displayed by a chi-squared four field test, which has in this case, a hit rate of 79%

Hope this helps you and your partner.

greetings

boris

Looks great, but in real life you will not work with 3 dimensions and 2 char values in each of them π

Fantastic and clever use of scenario to show the power of technology!

nice attempt....

Hi Boris

nice analysis. Well done!

greetings

cherry

its really nice.. well done Boris !!!

Nice one...

Good Analysis...

Best Regards,

Naresh K.

Great blog! Thanks