Advertisement
Guest User

Untitled

a guest
Aug 17th, 2019
127
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 2.11 KB | None | 0 0
  1. # Mastery Data Challenge
  2. We'd like to get a sense of your Python programming ability and how you approach machine learning problems.
  3. This challenge will culminate in a video conversation to discuss your thought process and trade-offs.
  4. **There is no time limit but please do not spend more than a few hours. We understand this is a side exercise and may be partially complete.**
  5. Note, that Mastery is an applied machine learning team; thus, we heavily rely on open-source libraries and do not implement algorithms from scratch.
  6. Document your thought process, how to set up, run, and test your solution with the aim of reproducibility
  7.  
  8.  
  9. ## Section 1: Python Programming
  10. TBD but probably a quick debugging exercise
  11.  
  12. ## Section 2: Machine Learning Problems
  13. Please select one of the following problems to complete.
  14. ### Option A: Email Parsing & Interpretation
  15. Carrier representatives, who work to develop relationships with trucking companies, receive many emails daily from their contacts providiing information on truck availability.
  16. They currently interpret the information received in these emails and manually enter the data into their transportation management system (TMS).
  17. We're aim to make their lives easier by doing this data entry automatically for them!
  18. Develop a solution that interprets the content of the emails in order to understand at a minimum truck origin and available date.
  19. Additional fields that can be extracted from the emails include preferred truck destination, equipment type, and any information to identify the carrier.
  20. ### Option B: Freight Matching
  21. A core functionality of any transportation management system (TMS) is to provide a freight matching service; that is, matching a load to trucks available to transport it.
  22. Your task is to generate matches for all available loads.
  23. A dataset of historical loads, available trucks, and available loads are provided to assist you.
  24. ### Option C: Lane Demand Forecasting
  25. A lane represents a unique origin and destination (e.g. Chicago to Los Angeles).
  26. Given historial truck loads by lane, generate a forecast of number of monthly truckloads for each lane for the upcoming three months.
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement