CA-HVF2017: California Household Vehicle Fleet Dataset (2017)

A statewide synthetic vehicle fleet dataset containing 13 million households and over ~25+ million vehicles across California. This dataset provides detailed household-level vehicle ownership information, including vehicle type, powertrain, vintage, and body type, generated using a Multiple Discrete Continuous Extreme Value (MDCEV) model and a geographically explicit synthetic population.

This dataset is designed to support transportation modeling, energy/emissions analysis, policy scenario evaluation, EV adoption studies, and agent-based simulations.

---

Overview

Modern transportation models require realistic household vehicle fleets, but privacy constraints limit access to micro-level vehicle ownership data.
CA-HVF2017 fills this gap by providing a statewide, publicly available, synthetic, validated representation of vehicle ownership across all California census tracts.

The dataset includes:

• Household characteristics (demographics, size, income)
• Detailed vehicle fleets for each household
• Geographic attributes at the census-tract level
• Location-based accessibility and built environment indicators

All values are synthetic but statistically consistent with real-world data.

---

Data Dictionary: Input

Below is the full set of household-level variables included in the CA-HVF2017 dataset.

---

Household Demographics

Variable	Description	Values
child	Household has at least one child	0, 1
HOUSEID	Household ID	numeric
HHSIZE	Household size	numeric
HHSIZE1	Household size = 1	0, 1
HHSIZE2	Household size = 2	0, 1
HHSIZE3	Household size = 3	0, 1
HHSIZE4	Household size = 4 or more	0, 1
NUMADLT	Number of adults	numeric
NUMCHILD	Number of children	numeric
NUM_WORKERS	Number of workers in household	numeric
retired	Household has at least one retiree	0, 1

---

Race / Ethnicity Indicators

Variable	Description	Values
hhwhite	Householder identifies as White	0, 1
hhasian	Householder identifies as Asian	0, 1
hhblack	Householder identifies as Black or African American	0, 1
hhothers	Householder identifies as another race (not White/Black/Asian)	0, 1

---

Income Categories

Variable	Description	Values
income1	Income < $25,000	0, 1
income2	$25,000 ≤ income < $50,000	0, 1
income3	$50,000 ≤ income < $75,000	0, 1
income4	$75,000 ≤ income < $100,000	0, 1
income5	Income ≥ $100,000	0, 1

---

Life Cycle Categories

Variable	Description	Values
LIF_CYC1	1 adult, no children	0, 1
LIF_CYC2	2+ adults, no children	0, 1
LIF_CYC3	1 adult + child age 0–5	0, 1
LIF_CYC4	2+ adults + child age 0–5	0, 1
LIF_CYC5	1 adult + child age 6–15	0, 1
LIF_CYC6	2+ adults + child age 6–15	0, 1
LIF_CYC7	1 adult + child age 16–21	0, 1
LIF_CYC8	2+ adults + child age 16–21	0, 1
LIF_CYC9	Household has at least one senior (65+)	0, 1
LIF_CYC10	Household has 2+ seniors (65+)	0, 1

---

Housing & Tenure

Variable	Description	Values
hhown	Household owns home	0, 1
perrent	% rental housing in tract	numeric
perrent1	Rental housing < 25%	0, 1
perrent2	Rental housing 25–45%	0, 1
perrent3	Rental housing > 45%	0, 1

---

Work Status

Variable	Description	Values
work0	No members employed	0, 1
work1	1 worker in household	0, 1
work2	2 workers in household	0, 1
work3	3+ workers in household	0, 1

---

Geographic Identifiers

Variable	Description	Values
county_fips	County FIPS code	numeric
county_name	County name	character
state_fips	State FIPS code	numeric
state_name	State name	character
tractid	Census tract ID	numeric

---

Transit & Accessibility Variables

Variable	Description	Values
emp_zscore	Standardized jobs reachable by 30-min transit	numeric
tractmean	Average number of jobs reachable from tract	numeric
tas_acres	Total acres accessible via 30-minute transit	numeric
tci	Transit Connectivity Index (0–100)	0–100
hi_tps	AllTransit Performance Score ≥ 8	0, 1
transit_performance_score	Transit Performance Score (0–10)	0–10

---

Built Environment Variables

Variable	Description	Values
job_density	Jobs per km²	numeric
pop_density	People per km²	numeric
res_density	Housing units per acre (unprotected)	numeric
pct_ag_land	% agricultural land	numeric
pct_water	% water area	numeric
urban_cbsa	Census tract is urban	0, 1
walkndx	Walkability index (0–20)	0–20

---

Log-Transformed Built Environment / Transit Indicators

Variable	Description	Values
log_job_density	Log(job_density)	numeric
log_job_above8	Log(job_density) > 8	0, 1
log_job_below4	Log(job_density) < 4	0, 1
log_pop_density	Log(pop_density)	numeric
log_pop_above9	Log(pop_density) > 9	0, 1
log_pop_below3	Log(pop_density) < 3	0, 1
log_res_density	Log(res_density)	numeric
log_pct_agland	Log(pct_ag_land)	numeric
log_pct_water	Log(pct_water)	numeric
log_lastyear_zevpct	Log(previous-year ZEV share)	numeric

---

Zero-Emission Vehicle (ZEV) Exposure

Variable	Description	Values
lastyear_zev_pct	Percentage of ZEVs in prior year	numeric

Data Dictionary: Output

Household-Level Variables

Variable name	Description	Value
county_fips	County FIPS code	numeric
tractid	Census tract ID	numeric
HOUSEID	Household ID	numeric
nvehicles	Number of vehicle(s) owned by the household	numeric

Vehicle-Level Variables

Variable name	Description	Value
county_fips	County FIPS code	numeric
tractid	Census tract ID	numeric
HOUSEID	Household ID	numeric
VEHID	Vehicle ID associated with the household	numeric
bodytype	The vehicle's body type	car, van, suv (Sport Utility Vehicle), pickup (Light-duty pick-up truck)
vintage_category	Vehicle age range	0–5 years, 6–11 years, 12+ years
annual_mileage	The vehicle's annual mileage	numeric
pred_power	The vehicle's powertrain	ICE (Internal Combustion Engine), AEV (All-Electric Vehicle), PHEV (Plug-in Hybrid Electric Vehicle)
modelyear	Vehicle model year (year manufactured)	numeric

---

Methodology Summary

1. Synthetic Population

Generated using PopulationSim, producing approximately 13 million California households with demographics matched to ACS distributions.

2. Multiple Discrete-Choice Vehicle Ownership Model

The dataset extends the MDCEV-based fleet composition model by Garikapati et al. (2014).
The model jointly predicts:

• Number of vehicles per household
• Vehicle category combinations
• Powertrain shares
• Vintage distributions

Predictors include:

• Income, household size, and workers
• Built environment metrics
• Accessibility indices
• Regional land-use patterns

3. Validation

The synthetic fleet is externally validated against:

• California DMV vehicle registration data
• County-level vintage and powertrain distributions
• Household vehicle count statistics

The dataset reproduces observed distributions with high fidelity.

---

Geographic and Temporal Scope

• Region: California
• Spatial resolution: Census tract (GEOID)
• Households: ~13 million
• Vehicles: ~25+ million
• Base demographic year: 2017
• Fleet calibration year: 2017

---

License

This dataset is licensed under the Creative Commons Attribution 4.0 International (CC BY 4.0) license.

You are free to:

• Share — copy and redistribute the material in any medium or format
• Adapt — remix, transform, and build upon the material for any purpose, even commercially

Under the following terms:

• Attribution — You must give appropriate credit, provide a link to the license, and indicate if changes were made.
  You may do so in any reasonable manner, but not in a way that suggests the licensor endorses you or your use.

Full license text: https://creativecommons.org/licenses/by/4.0/

---

How to Load the Dataset

Python (pandas)

import polars as pl

hh = pl.read_parquet("households.parquet")
veh = pl.read_parquet("vehicles.parquet")