Skip to main content

Analyze Output Events

After running a simulation, an events.txt file is generated. This file records simulation events in JSON Lines format, where each line represents a single event encoded as a JSON object. Events are written in chronological order, reflecting the simulation timeline. The structure and semantics of each event are defined in the Event Specification.

Here we demonstrate simple analysis examples on the event log stored in events.txt, focusing on both user behavior and mobility unit usage, using Python and pandas

User-focused analysis​

The following code selects all non-walking RESERVED events in a DataFrame for further analysis.

import pandas as pd

# Read line-delimited JSON
df = pd.read_json("events.txt", lines=True)

# Extract RESERVED events
df = df[df["eventType"] == "RESERVED"]

# Exclude walking-based reservations
df = df[df["source"] != "walking"]

# Expand relevant fields from 'details'
df["userId"] = df["details"].apply(lambda d: d.get("userId"))
df["demandId"] = df["details"].apply(lambda d: d.get("demandId"))
df["success"] = df["details"].apply(lambda d: d.get("success"))
df["org"] = df["details"].apply(lambda d: d["route"][0]["org"]["locationId"] if d.get("route") else None)
df["dst"] = df["details"].apply(lambda d: d["route"][0]["dst"]["locationId"] if d.get("route") else None)

# Select and sort columns for easier analysis
df["service"] = df["source"]
df = df[["time", "userId", "demandId", "service", "org", "dst", "success"]]
df = df.sort_values(["userId", "time"])

Output:

             time  userId demandId service   org   dst  success
69     563.309169  CU_001      D_1    gtfs  1_01  7_01     True
2759  1120.356102  CU_001      D_2    gtfs  7_02  2_02     True

This table allows you to see:

  • Which services users reserved and whether the reservations succeeded
  • The origin and destination of each trip
  • How the timing and frequency of reservations relate to user demand

Mobility-focused analysis​

The following code selects only the segments where users actually rode a mobility unit in a DataFrame for further analysis.

import pandas as pd

# Read line-delimited JSON
df = pd.read_json("events.txt", lines=True)

# Keep only DEPARTED and ARRIVED events
df = df[df["eventType"].isin(["DEPARTED", "ARRIVED"])]

# Extract mobilityId and userId from details
df["mobilityId"] = df["details"].apply(lambda d: d.get("mobilityId"))
df["userId"] = df["details"].apply(lambda d: d.get("userId"))
df["demandId"] = df["details"].apply(lambda d: d.get("demandId"))
df["locationId"] = df["details"].apply(lambda d: d["location"].get("locationId") if d.get("location") else None)
df["time"] = df["time"]

# Keep only rows where both mobilityId and userId are present
df = df[df["mobilityId"].notna() & df["userId"].notna()]

# Select relevant columns
df = df[["time", "eventType", "userId", "mobilityId", "demandId", "locationId"]]

# Sort by userId, mobilityId, and time
df = df.sort_values(["userId", "mobilityId", "time"]).reset_index(drop=True)

Output:

     time eventType  userId               mobilityId demandId locationId
0   570.0  DEPARTED  CU_001  1平日・土日祝_09時30分_系統101001      D_1       1_01
1   578.0   ARRIVED  CU_001  1平日・土日祝_09時30分_系統101001      D_1       7_01
2  1121.0  DEPARTED  CU_001  1平日・土日祝_18時00分_系統101002      D_2       7_02
3  1131.0   ARRIVED  CU_001  1平日・土日祝_18時00分_系統101002      D_2       2_02

This table allows you to see:

  • Which users rode which mobility units and during which time intervals
  • The departure and arrival times for each passenger trip segment