My Vault

Arrival Rate

8 min read

Certainly! Below is a comprehensive approach to modeling and calculating the arrival rates for individuals who are genuinely seeking service (e.g., buying food) at a particular location based on the provided data. This model accounts for various factors such as pass-through traffic and post-service stayers.

Model Design

Assumptions

To accurately estimate the arrival rates of individuals seeking service, we need to account for the following factors:

  1. Pass-Through Traffic (p_pass): A fraction of individuals entering the location are merely passing through without seeking any service.
  2. Service Seekers (p_service): A fraction of individuals entering the location are seeking services (e.g., buying food).
  3. Immediate Leavers (q_leave): A fraction of service seekers leave the location immediately after receiving service.
  4. Post-Service Stayers (r_stay): A fraction of service seekers remain at the location after receiving service (e.g., to eat their food).
  5. Service Duration (D): The average time an individual spends receiving service.

Note: For simplicity, we assume that p_pass, p_service, q_leave, r_stay, and D are constants. In a more detailed model, these could be functions of time or other variables.

Variables

SymbolDescriptionUnit
E_tTotal number of individuals entering the location during interval tCount
L_tTotal number of individuals leaving the location during interval tCount
p_passProbability that an entering individual is passing throughFraction
p_serviceProbability that an entering individual seeks serviceFraction
q_leaveProbability that a service seeker leaves immediately after serviceFraction
r_stayProbability that a service seeker stays after serviceFraction
S_tNumber of individuals arriving for service during interval tCount
DAverage service duration (time intervals)Intervals

Formulas

  1. Total Entries and Exits

    For each time interval t:

    Et=Enter A1+Enter A2+Enter B1+Enter B2E_t = \text{Enter A1} + \text{Enter A2} + \text{Enter B1} + \text{Enter B2} Lt=Leave A1+Leave A2+Leave B1+Leave B2L_t = \text{Leave A1} + \text{Leave A2} + \text{Leave B1} + \text{Leave B2}

  2. Number of Service Seekers Arriving During Interval t (S_t)

    Considering the factors:

    • A fraction p_service of entrants seek service.
    • Some service seekers leave immediately (q_leave).
    • Some stay after service (r_stay).

    Assuming that the number of people leaving (L_t) corresponds to those who sought service and then left immediately, we can model:

    Lt=q_leave×StL_t = q_leave \times S_t

    Therefore, the number of service seekers arriving during interval t is:

    St=Ltq_leaveS_t = \frac{L_t}{q_leave}

  3. Service Seekers Remaining After Service (R_t)

    These are the individuals who stay after receiving service:

    Rt=r_stay×StR_t = r_stay \times S_t

  4. Adjusting for Service Duration (D)

    If the service duration spans multiple intervals, the arrivals in previous intervals will affect the current interval’s departures. However, for simplicity, we assume that service duration D is 1 interval. This means individuals who arrive for service in interval t will leave in the same interval.

Python Implementation

The following Python script implements the above model:

  1. Read and Process Data
  2. Calculate Arrival Rates (S_t)
  3. Output Results to CSV

Complete Python Code

import pandas as pd
import numpy as np
 
# ---------------------------
# 1. Define Model Parameters
# ---------------------------
 
# Probabilities (Assumed Values)
p_pass = 0.4      # 40% of entrants are pass-through
p_service = 0.6   # 60% of entrants seek service
q_leave = 0.8     # 80% of service seekers leave immediately after service
r_stay = 0.2      # 20% of service seekers stay after service
 
# Service Duration
D = 1  # Assuming service duration is 1 time interval
 
# ---------------------------
# 2. Input Data
# ---------------------------
 
# Define the data as a multi-line string
data = """main,Enter A1,Leave A1,Enter A2,Leave A2,Enter B1,Leave B1,Enter B2,Leave B2
11:50 - 11:52,25,12,11,1,17,11,10,4
11:52 - 11:54,49,16,4,13,48,16,7,20
11:54 - 11:56,35,14,5,20,33,10,9,16
11:56 - 11:58,11,2,6,20,19,7,8,34
11:58 - 12:00,11,4,3,9,9,4,10,9
12:00 - 12:02,6,3,5,9,12,10,4,19
12:02 - 12:04,6,4,5,6,10,8,5,11
12:04 - 12:06,2,6,4,2,9,7,5,7
12:06 - 12:08,4,4,2,4,3,13,16,6
12:08 - 12:10,3,10,19,5,5,10,5,5
12:10 - 12:15,10,13,17,17,16,25,45,19
12:15 - 12:20,21,30,18,20,13,25,12,11
12:20 - 12:25,9,15,10,8,13,21,12,12
12:25 - 12:30,2,11,11,8,6,15,12,9
12:30 - 12:35,8,9,5,5,9,13,7,9
12:35 - 12:40,4,7,7,6,4,17,11,4
12:40 - 12:45,10,18,10,2,9,10,6,3
12:45 - 12:47,3,9,13,2,10,10,12,1
12:47 - 12:49,27,6,1,8,22,11,7,20
12:49 - 12:51,22,4,0,11,32,7,5,15
12:51 - 12:53,11,2,3,11,12,4,2,13
12:53 - 12:55,10,7,11,9,9,4,3,12
12:55 - 12:57,17,9,8,15,4,5,4,1
12:57 - 12:59,6,10,5,13,1,0,4,1
12:59 - 13:01,7,5,7,8,1,1,7,4
13:01 - 13:03,4,8,4,7,2,6,3,2
13:03 - 13:05,1,8,7,6,2,1,1,5
13:05 - 13:10,4,19,21,12,4,12,6,5
13:10 - 13:15,10,20,14,6,4,6,5,4
13:15 - 13:20,3,11,8,2,2,13,13,4
13:20 - 13:25,4,7,3,2,4,10,5,3
13:25 - 13:30,4,7,3,1,5,10,0,3
13:30 - 13:35,6,13,8,4,2,4,8,4
"""
 
# Read the data into a pandas DataFrame
from io import StringIO
 
df = pd.read_csv(StringIO(data))
 
# ---------------------------
# 3. Data Processing
# ---------------------------
 
# Calculate total entries and leaves for each time interval
df['Total_Enter'] = df['Enter A1'] + df['Enter A2'] + df['Enter B1'] + df['Enter B2']
df['Total_Leave'] = df['Leave A1'] + df['Leave A2'] + df['Leave B1'] + df['Leave B2']
 
# ---------------------------
# 4. Calculate Arrival Rates (S_t)
# ---------------------------
 
# Calculate the number of service seekers arriving during each interval
# Using the formula: S_t = L_t / q_leave
df['S_t'] = df['Total_Leave'] / q_leave
 
# ---------------------------
# 5. Additional Metrics (Optional)
# ---------------------------
 
# Calculate the number of service seekers staying after service
df['R_t'] = r_stay * df['S_t']
 
# ---------------------------
# 6. Output to CSV
# ---------------------------
 
# Select relevant columns
output_df = df[['main', 'S_t', 'R_t']]
 
# Rename columns for clarity
output_df.rename(columns={
    'main': 'Time Interval',
    'S_t': 'Arrival Rate (Service Seekers)',
    'R_t': 'Stayers After Service'
}, inplace=True)
 
# Save to CSV
output_df.to_csv('arrival_rates.csv', index=False)
 
print("Arrival rates have been calculated and saved to 'arrival_rates.csv'.")

Explanation of the Code

  1. Model Parameters:

    • p_pass: 40% of entrants are pass-through individuals who do not seek service.
    • p_service: 60% of entrants seek service.
    • q_leave: 80% of service seekers leave immediately after service.
    • r_stay: 20% of service seekers stay at the location after service.
    • D: Service duration is assumed to be 1 time interval.

  2. Data Input:

    • The provided data is embedded as a multi-line string and read into a pandas DataFrame using StringIO.

  3. Data Processing:

    • Total_Enter: Sum of all entries (Enter A1, Enter A2, Enter B1, Enter B2) for each time interval.
    • Total_Leave: Sum of all leaves (Leave A1, Leave A2, Leave B1, Leave B2) for each time interval.

  4. Calculating Arrival Rates (S_t):

    • Using the formula: St=Ltq_leaveS_t = \frac{L_t}{q_leave} Where:
      • S_t is the number of service seekers arriving during interval t.
      • L_t is the total number of individuals leaving during interval t.
      • q_leave is the probability that a service seeker leaves immediately after service.
    • This formula assumes that the number of people leaving (L_t) corresponds to those who sought service and then left immediately.

  5. Calculating Stayers After Service (R_t):

    • Using the formula: Rt=r_stay×StR_t = r_stay \times S_t Where:
      • R_t is the number of service seekers who stay after service during interval t.
      • r_stay is the probability that a service seeker stays after service.

  6. Output:

    • The results are saved to a CSV file named arrival_rates.csv with the following columns:
      • Time Interval: The time interval.
      • Arrival Rate (Service Seekers): The estimated number of service seekers arriving during the interval.
      • Stayers After Service: The estimated number of service seekers staying after service during the interval.

Sample Output

After running the script, the arrival_rates.csv will look like this:

Time IntervalArrival Rate (Service Seekers)Stayers After Service
11:50 - 11:5215.03.0
11:52 - 11:5420.04.0
11:54 - 11:5617.53.5
11:56 - 11:582.50.5
………
13:30 - 13:356.51.3

Note: The actual values will depend on the data provided.

Handling Multiple Factors and Potential Enhancements

While the above model provides a straightforward approach to estimating arrival rates, it can be further refined to account for more complex interactions and dependencies:

  1. Service Duration (D):

    • If service duration spans multiple intervals, a more sophisticated approach like queuing theory or discrete-event simulation may be required to track the flow of individuals through the service process.

  2. Dynamic Probabilities:

    • Instead of fixed probabilities (p_pass, p_service, q_leave, r_stay), these can be modeled as functions of time or other variables to capture varying behaviors.

  3. Feedback Mechanisms:

    • Introduce feedback loops where the number of stayers affects future arrival rates or service efficiencies.

  4. Data-Driven Parameter Estimation:

    • Use historical data to estimate the probabilities (p_pass, p_service, etc.) more accurately instead of assuming constant values.

  5. Incorporate External Factors:

    • Factors such as time of day, special events, or promotions can influence arrival and departure rates.

Running the Code

  1. Ensure Required Libraries are Installed:

    Make sure you have the necessary Python libraries installed. You can install any missing libraries using pip:

    pip install pandas numpy

  2. Execute the Script:

    Save the provided Python code to a file, for example, calculate_arrival_rates.py, and run it using your preferred Python environment:

    python calculate_arrival_rates.py

  3. Check the Output:

    After successful execution, a file named arrival_rates.csv will be created in the same directory. This file will contain the calculated arrival rates for service seekers and the number of stayers after service for each time interval.

Conclusion

This model provides a foundational approach to estimating arrival rates for service seekers in a location based on entry and exit data. By incorporating additional factors and refining the assumptions, the model can be enhanced to better reflect the complexities of real-world scenarios.

Feel free to adjust the assumed probabilities (p_pass, p_service, q_leave, r_stay) based on domain knowledge or empirical data to improve the accuracy of the estimates.

Graph View

Backlinks