Party Data Window Computation | Dataford Interview Questions

Given JSON party submission data (start_time, end_time, party_id) and geographic data (neighborhood, town, state, city, party_id), compute party_windows: for each neighborhood, the (start_hour, end_hour) windows where at least one party has begun and at least one has not yet ended.

Asked in the technical phone screen stage. Prep guidance mentioned reviewing strptime and datetime.

Implement compute_party_windows(submissions, geography). Both inputs are lists of dictionaries, and times use %Y-%m-%d %H:%M:%S on the same date. Treat each party interval as active in every intersecting hourly bucket, with the end hour exclusive. Return a dictionary mapping each neighborhood to sorted, merged [start_hour, end_hour] windows. Include neighborhoods with no matching parties as empty lists.

Constraints

0 <= len(submissions), len(geography) <= 10^5
Times use the format %Y-%m-%d %H:%M:%S
All timestamps occur on the same calendar date
Every submission has start_time < end_time
Party IDs are unique within each input
A submission whose party ID is absent from geography is ignored

Function Signature

def compute_party_windows(submissions, geography):

Your question is Party Data Window Computation. Start with the requirements in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.

Problem

Asked in the technical phone screen stage. Prep guidance mentioned reviewing strptime and datetime.

Constraints

0 <= len(submissions), len(geography) <= 10^5
Times use the format %Y-%m-%d %H:%M:%S
All timestamps occur on the same calendar date
Every submission has start_time < end_time
Party IDs are unique within each input
A submission whose party ID is absent from geography is ignored

Function Signature

def compute_party_windows(submissions, geography):

Your question is Party Data Window Computation. Start with the requirements in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.