You will be given dataset descriptions and must identify forecasting problems that could be solved using each dataset. For each dataset, identify at least 2 different forecasting problems and evaluate their feasibility.
DATASET A: Retail Chain Sales Data
3 years of weekly sales data from 150 store locations
Variables: Store ID, week, total revenue, units sold by product category (electronics, clothing, home goods), local weather data, local economic indicators, holiday flags, promotional activity flags Granularity: Weekly totals per store Time period: January 2021 - December 2023
Forecasting Problem #1:
Forecasting Problem #2:
Data Sufficiency Check:
DATASET B: University Enrollment Records
5 years of semester enrollment data across all programs
Variables: Semester, program/major, new enrollments, total enrollments, graduations, demographic breakdowns (age, geography), application numbers, admission rates, tuition fees, employment rates of graduates Granularity: Semester totals by program Time period: Fall 2019 - Spring 2024
Forecasting Problem #1:
Forecasting Problem #2:
Data Sufficiency Check:
DATASET C: City Transportation Usage
2 years of public transportation ridership data
Variables: Date, bus route, train line, daily passenger counts, weather conditions, special events, fuel costs, service disruptions, time of day usage patterns, seasonal tourism data Granularity: Daily totals by transportation mode and route Time period: March 2022 - February 2024
Forecasting Problem #1:
Forecasting Problem #2:
Data Sufficiency Check:
DATASET D: Hospital Patient Flow
18 months of daily hospital operations data
Variables: Date, daily admissions, daily discharges, emergency room visits, surgical procedures scheduled, bed occupancy rates, staffing levels, seasonal illness patterns, local population demographics Granularity: Daily hospital-wide totals Time period: January 2023 - June 2024
Forecasting Problem #1:
Forecasting Problem #2:
Data Sufficiency Check:
Critical Thinking Questions (5 minutes)
Choose your strongest forecasting problem from above and answer:
1. Validation: How would you test if your forecast is working?
2. Stakeholder Impact: Who else in the organization would be affected by decisions based on this forecast?
3. Risk Assessment: What happens if your forecast is significantly wrong?
4. Alternative Approaches: Could this problem be solved with prediction methods instead? Why or why not?