Solutions to "Just KNIME It!" Challenge 3

Just KNIME It!
21

Link updated : KNIME_Challenge-3 – KNIME Hub

22

The dataset is messed up. In fact, “Prostate” is the most frequent male cancer site in 2017, but the whole “Male Genital System” section is not present in the dataset, so “Prostate” does not rank in the top 5 sites.

23

Here is my solution. I get a lot out of this (KNIME it) by 1) make my own solution 2) look at other solutions 3) see the differences and learn.

Capture
Capture1302×617 78.6 KB

1 Like
24

Hi everyone,
Here is my solution.

Just Knime it - 3

1 Like
25

Hello KNIMErs,

Here is my solution to #justknimeit-3 :

KNIME Hub > gonhaddock > Spaces > Just_KNIME_It > Just KNIME It _ Challenge 003

image
image958×888 70.2 KB

Female: GREEN
Men: RED

BR

4 Likes
26

I’ve tried to correct the dataset. This is my solution
KNIME_challenge_3.knwf (69.6 KB)

2 Likes
27

a 2nd versión, working on database in-memory

image
image475×659 42.5 KB

In-memory database

4 Likes
28

After reading some of other solutions I noticed that I missed it out that the data were mixed with the totals , I think now it is working fine

We should exclude the true in bellow

image

1 Like
29

Hello,
I did some updates and modifications to the workflow. Some literature has been added order to clarify my results as well:

  1. An error on 'Cancer Site Code’s Rule Engine has been amended.
  2. It has been connected the exclusion of ‘All invasive Cancer Sites Combined’ site type, aiming to avoid bias in final results.

image
image1540×500 57.8 KB

image
image924×603 22.3 KB

Female: GREEN
Male: RED

BR

3 Likes
31

My solution on Knime Hub KNIME_challenge_3 – KNIME Hub

2 Likes
32

Hi,
here my solution: knime://My-KNIME-Hub/Users/cf_123/Public/jKi3

image
image822×489 22.5 KB

1 Like
33

I’m (finally) not the author of this week’s challenge, so I can participate and share my proposed solution!

Thank you guys for pointing out your assumptions and concerns in this thread. I have also included mine (i.e. I excluded some cancer sites since they are “aggregations” of other counts).

I also got inspired by the picture in the challenge page and came up with a Choropleth map showing the incidence rate by state.

image
image852×527 58.3 KB

To do that, I used the Choropleth World Map component available on the KNIME Hub and modified the JavaScript code adding the option resolution:“provinces”.

9 Likes
34

here is my solution for this week’s challenge

I grouped the different breast cancer types distinguishing between male and female into one category and estimated the missing values as group means, based on gender & age to get an accurate estimate

3 Likes
35

Hi all,

Here is my solution :slight_smile: to this challenge and my very first contribution on the KNIME Hub!

4 Likes
36

Hi @duristef,

Thanks for your contribution!

Yes, as you’re hinting at, the results are going to be as good as the data is. It turns out that the data has problems that are going to end up reflecting on the solution. Our main goal here, however, is on how to answer the frequency questions by processing the data. How to clean the data, or how to improve its quality, are secondary (and more advanced) goals.

Cleaning and fixing datasets is definitely paramount for serious data science, and although we had not set out to focus on this here, we are pleased to see that our community is paying attention to it. :slight_smile:

5 Likes
37

Here’s our team’s solution to the challenge!

As mentioned yesterday, results may vary depending on how you process or interpret the data. A detailed explanation to our solution is here: Just KNIME It! | KNIME

We’re very happy with the diversity in the answers! Big shout-out to everybody who participated and see you tomorrow with a new challenge.

#justknimeit #justknimeit-3

4 Likes
38

Hi! Here is my solution :slight_smile:

Challenge 3 - Just KNIME It!

2 Likes