Titanic Roster Project


Home News About data:Train data:Test data:Real

------------- News ------------
This page simply has links to make it easy for you to find the latest news or video.
If you have information you would like to see added to this page, please email me at vbiersch@gmail.com

------------- 3 Sets of Data ------------
The Train Data is used to train for this exercise
The Test Data is the data that will be submitted once I have developed the programs to correctly calculate who will survive based on the Train Data.
The Real Data will be where I can make it easier for people to view, sort, print out the real data to help them in their own research once I find a good source of data.
------------- My Background -------------

Developed by a Navy Veteran with over 30 years of high tech experience working as a Navy Radioman, an electronics tech, and wearing all hats in software development.

It seems that via age discrimination, caste and neopotism, my skills are no longer needed in corporate America.

So I sought work in government service with organizations like the Department of Veterans Affairs only to find that they too have chosen to hire non-immigrant guest workers on visas like the H-1B instead of veterans like myself.

So I sought entry level jobs at the nearest VA hospital hoping to be given the opportunity to work my way up the ladder only to quit in disgust after 4 attempts when I was told I would need to ask for permission to go to the bathroom.

Bottom line, I need work!

If you would like to hire me for your current project, I am available.

Better yet, if you would like to hire me as a employee where I could count on having a job until I can't work no more, I would appreciate it.

Did I tell you that since I was forced to take early social security, I can work cheap?

------------- Why This Project -------------

I've been hearing a lot about machine learning so I was googling it, and ran across this project.

https://www.kaggle.com/c/titanic/overview

Being an Amateur Radio Operator (N5IVV) and a Navy Veteran, I found it interesting.

After looking at their recommended tools and samples of the software, I said, I bet I can do it with old tools like PHP, and since I can't find work as an older software developer, I decided to do it using PHP/MySQL and throw in some newer tools like highcharts, .datatable.net and the google maps javascript API simply because I wanted to show that older tools are quite capable of producing the same end product.

And here we are.

This site will be a work in progress as I think of different ways to solve their problem.

I hope you enjoy it, and please tell your friends about it!

------------- My Goal -------------

I decided to use PHP just to show that it is not the tool, but the knowledge that the skilled worker brings to the table that makes a difference and ask a simple question.
Why are we throwing away our highly skilled workers by sending their jobs offshore or importing non-immigrant guest workers to take their place?

------------- My Submissions -------------

First Pass on 22 Oct 2020
Received a score of .71291
My code for the first entry can be viewed by Clicking Here
You can Click Here to view all of the scores and where I stood on this date
I have highlighted the ones that had 3 digit scores in blue so that you can identify this sweet spot I'm talking about

Second Pass on 23 Oct 2020
Received a score of .57177
This was surprising in a way but in another way I expected that the only way to get 100% is to use real live data and submit a spreadsheet with those results
But I'm an ethical type guy so I will post my matching percentages spreadsheet here so you can see what I'm talking about.
Click Here to view the spreadsheet picture.
My code for the second entry can be viewed by clicking Here
You can Click Here to view all of the scores and where I stood on this date

Third Pass on 23 Oct 2020
Received a score of .71770
My code for the third entry can be viewed by clicking here
Click Here to view the spreadsheet picture.

Fourth Pass on 24 Oct 2020
Received a score of .73923
My code for the fourth entry can be viewed by clicking here
Click Here to view the spreadsheet picture.
I think if I round up instead of down in the code I can increase the score a bit if you look at the highlighted cells on the spreadsheet.

Fifth Pass on 24 Oct 2020
Received a score of .72966 after rounding up in the previous submission
At this point, it would require me creating a table to save the data I submitted each time and then processing that data each time in various combinations
Unfortunately, in desperation, I had to accept a job stocking grocery store shelves for low pay, and I start on monday so that is not possible
I believe if you only use data and the percentages, I have taken it as high as possible by focusing on the lower level detail in the code
To take it further would require developing the database matrix I discussed above and guessing, or comparing real live data and manually preparing the submission file to achieve a perfect score
And I'm just not going to go there