Documentation

Verified package

File Structure

The current public sample archive contains the following files.

complete_trip_saltlake_2020_1day_Sample_v1.0/
├── data/
│   └── part-001-of-001.parquet
├── data_dictionary.csv
└── Getting_Started.ipynb

Hands-on analysis

Example Analysis Notebook

The example notebook demonstrates how to locate and load the public sample, inspect complete trips and travel segments, run quality checks, summarize travel modes, link related records, and use the companion network tables to visualize a selected route without exposing unnecessary identifiers.

from pathlib import Path
import pandas as pd

data_dir = Path("data")
csv_files = list(data_dir.glob("*.csv"))
parquet_files = list(data_dir.glob("*.parquet"))

def load_table(path: Path) -> pd.DataFrame:
    if path.suffix.lower() == ".csv":
        return pd.read_csv(path)
    if path.suffix.lower() == ".parquet":
        return pd.read_parquet(path)
    raise ValueError(f"Unsupported: {path.suffix}")

files = csv_files or parquet_files
if not files:
    raise FileNotFoundError(
        "Download and extract the public sample first."
    )

table = load_table(files[0])
print(table.shape)
print(table.columns.tolist())
table.head()

missing_summary = (
    table.isna()
    .sum()
    .sort_values(ascending=False)
    .rename("missing_values")
)
missing_summary.head(20)

candidate_trip_id_columns = [
    c for c in table.columns
    if "trip" in c.lower() and "id" in c.lower()
]

candidate_coordinate_columns = [
    c for c in table.columns
    if any(term in c.lower() for term in
           ["lat", "lon", "lng", "latitude", "longitude"])
]

Interpretation

Known Limitations

• The public sample covers one undisclosed day and may not represent all seasonal or long-term travel patterns.
• Spatial and temporal information may be aggregated or generalized for privacy.
• Travel modes are inferred and may contain classification uncertainty.
• Device-based mobility observations may not represent the full population.
• Some complete trips may contain multiple travel modes.
• The public sample is intended for evaluation and reproducibility support, not as a replacement for the controlled full dataset.