Enis G.

Enis G.

B.Sc. & M.Sc. in Statistics, PhD (c), Data Scientist & DevOps, Regular Kaggler, AWS Community Builder, 2x AWS Certified

Setur Datathon - 2nd Place Prize Winner

Published:

alt text

TR tr

  • Yarışma hakkında detaylı blog içeriğine buradan ulaşabilirsiniz.


ENG uk

  • Team Name: Runtime Terror
  • Team Members: Enis Gumustas & Alper Kocabiyik & Hamza Tuna
  • Mainly we built a recommendation engine “Gezgin” for a tourism agency (Setur). Gezgin is working with two different strategies as user-based and cluster-based and serving with an API.
  • We calculated an affinity score for the customers using profiles, segments and hotel-specific features that we created from various tables for hotel, tour and cruise recommendations.
  • We developed a recommendation engine by using the calculated affinity score with ALS.
  • We have also developed a clustering-based recommendation engine for variables that are not available in tour and cruise.
  • In the case of clustering-based customers who do not have any purchase history, we have created a solution to the Cold Start problem that is frequently encountered in recommendation engines.
  • Technologies used: PySpark, Pandas, MongoDB, Flask.

Setur Datathon