California COVID-19 Vaccines Administered By County

Rows: 49,538
Columns: 13
Data range: 2024-08-21 to 2026-10-28
Auto-refresh: Re-checks its source daily
Last refreshed with new data: 2026-10-30

Daily time-series COVID-19 vaccination administration data by California county including total doses, cumulative doses, partially and fully vaccinated counts. Historical and recent data.

AI analysis

Data Quality & Structure Analysis: California COVID-19 Vaccines by County

1. Dataset Overview and Value

This dataset provides daily time-series COVID-19 vaccination administration figures across California's 58 counties (plus statewide/aggregate rows, given 62 distinct COUNTY values), sourced from the California Health and Human Services (CHHS) open data portal (data.chhs.ca.gov). The CSV-based API delivery suggests a stable, machine-readable pipeline suitable for automated ingestion. Notably, the time coverage window (2024-08-21 through 2026-07-04) is unusual for a "COVID-19 vaccination" dataset — this either reflects a rolling/forward-dated publishing schedule, placeholder future dates, or a system clock artifact in the source; this should be verified before treating the recency as a strict signal of freshness. With 42,346 rows spanning daily granularity across counties and 683 distinct dates, this dataset is well-suited for time-series and geographic correlation studies — e.g., examining vaccination uptake against case rates, hospitalization trends, policy interventions, or demographic/socioeconomic factors at the county level.

2. Data Quality Observations

Overall null density is low: 6,812 null cells out of a possible ~550,498 (42,346 rows × 13 columns), or roughly 1.2% of all cells. However, this null count is entirely concentrated in one column — CALIFORNIAFLAG, which is missing in 16.1% of rows (6,812/42,346). This is a substantial and disproportionate gap for a single field and suggests the flag may only be populated for aggregate/statewide rows (or vice versa) rather than being a true missing-data problem — this should be investigated as a structural/conditional null rather than random data loss before excluding rows. All other 12 columns report zero nulls, which is a strong quality signal for the core numeric measures. Duplicate-row counts are not yet finalized ("Phase D — recomputed against Parquet"), so deduplication should be confirmed before running aggregations, since daily county-level time series are especially vulnerable to duplicate-date/county pairs inflating totals.

3. Key Columns and Distributions

The core dose-count columns are heavily right-skewed, consistent with county population disparities (e.g., Los Angeles vs. small rural counties) and reporting-period spikes. TOTALDOSES has a mean of 800 vs. median of 28 (skew=12, 6,540 outliers), indicating a small number of high-volume days/counties dominate the distribution — Q1=3, Q3=187 shows the bulk of daily records are very low-volume. CUMULATIVETOTALDOSES ranges from 1,732 to ~108.6 million with a similarly large skew (5) and gap between median (462,156) and mean (~5.03M), reflecting the wide population/time-accumulation range across counties. FULLYVACCINATED (daily, not cumulative) is nearly degenerate — median and Q3 both 0, max only 41, skew=16.4 — suggesting this metric is now rarely updated or near end-of-life in the source data (consistent with declining reporting cadence in 2024+). UPTODATECOUNT and its cumulative counterpart also show heavy right-skew (16.2 and 8.61) with medians of 0, indicating that "up to date" booster-style metrics are sparse/zero for most county-day records but spike significantly in a subset (8,279 and 8,363 outliers respectively) — likely tied to specific reporting counties or campaign periods.

4. Recommended Join Keys

- ADMINISTEREDDATE and COUNTY together form the natural composite key for this dataset (683 distinct dates × 62 distinct counties). - For cross-dataset correlation, COUNTY is the strongest candidate join key against other California county-level datasets (case counts, hospitalizations, mortality, demographics, mobility). - ADMINISTEREDDATE should be normalized to a standard date grain (daily) to align with other CHHS time-series datasets, which typically share this same date convention.

5. Suggested Complementary Datasets

Strong pairing candidates include: California COVID-19 case rates and hospitalizations by county (also CHHS-published, likely sharing the same COUNTY/date schema) to correlate vaccination uptake with subsequent case/hospitalization decline; county-level demographic and socioeconomic data (Census/ACS) to examine equity gaps in vaccination coverage; CDC or state booster/up-to-date vaccination datasets to cross-validate the UPTODATECOUNT anomalies; and mobility or policy-intervention datasets (e.g., Google/Apple mobility, county mask/vaccine mandate timelines) to study behavioral correlations with vaccination campaign timing.

Columns

ADMINISTERED_DATE (date)
COUNTY (string)
TOTAL_DOSES (integer)
CUMULATIVE_TOTAL_DOSES (integer)
PARTIALLY_VACCINATED (integer)
TOTAL_PARTIALLY_VACCINATED (integer)
FULLY_VACCINATED (integer)
CUMULATIVE_FULLY_VACCINATED (integer)
AT_LEAST_ONE_DOSE (integer)
CUMULATIVE_AT_LEAST_ONE_DOSE (integer)
CALIFORNIA_FLAG (string)
UP_TO_DATE_COUNT (integer)
CUMULATIVE_UP_TO_DATE_COUNT (integer)