0 votes

Hi Team,

I have created a simple recommendation system in a python notebook. I started by changing the dataiku.dataset to a dictionary.

This is the final function:

Here you can see the results for family nr. 13. 

However, I would like to get recommendations for all family-numbers. The output should be five columns. The first column holding the top recommendation for each family, the second column holding the next best recommendation, etc

asked by
Hi,
Are you getting any errors specific to Dataiku?
Otherwise, this is more a Python question.  For data wrangling purposes, we recommend using pandas, which has useful groupby and apply methods for dealing with such use cases.
Hope it helps,
Alex
Hi, have you been able to implement a code solution for your case?
Yes, this has been solved. Thanks for your help.
Great. Would you be OK to add your solution to this thread as an answer? This would help others in the Dataiku community :) Otherwise we can close this thread.
Hi, Gentle reminder to add the solution you found to help the community. Thanks!

1 Answer

0 votes

Hi. In the end I had correctly coded the transformation to a dataframe. The issue was that it was running for hours. We solved it with multi-processing, which only took a fraction of the time. 

answered by
thanks!
994 questions
1,023 answers
1,075 comments
3,027 users

┬ęDataiku 2012-2018 - Privacy Policy