How to Scrape US Census Data: Complete Guide for 2026

This guide covers extracting publicly accessible data. Always review a site's robots.txt and Terms of Service before scraping.

TL;DR

To scrape US Census data in 2026, use AlterLab's API with automatic tier escalation. Start with T1 for static pages, then let the API promote to T3/T4 as needed for anti-bot protections. Extract structured data via CSS selectors or Cortex AI for typed JSON output.

Why collect government data from US Census?

Government data provides reliable foundations for analysis. Common use cases include:

• Market researchers tracking demographic shifts for product expansion planning
• Analysts correlating housing statistics with economic indicators
• Developers building public policy tools that require authoritative population data

Technical challenges

US Census implements standard anti-bot measures: rate limiting by IP, User-Agent validation, and JavaScript challenges on dynamic pages. Raw HTTP requests often fail with 403 or 429 responses. AlterLab's Smart Rendering API automatically handles these via rotating residential proxies, realistic browser fingerprints, and tiered rendering approaches—escalating only when necessary to avoid overpaying for simple pages.

Quick start with AlterLab API

See the Getting started guide for SDK installation. Below are examples for scraping a public Census page showing median household income data.

Python example:

import alterlab

client = alterlab.Client("YOUR_API_KEY")
response = client.scrape("https://www.census.gov/programs-surveys/acs/")
print(response.text)

Node.js example (MUST include this):

import { AlterLab } from "alterlab";

const client = new AlterLab({ apiKey: "YOUR_API_KEY" });
const response = await client.scrape("https://www.census.gov/programs-surveys/acs/");
console.log(response.text);

cURL example:

curl -X POST https://api.alterlab.io/v1/scrape \
  -H "X-API-Key: YOUR_KEY" \
  -d '{"url": "https://www.census.gov/programs-surveys/acs/"}'

Extracting structured data

For the ACS program page, target these visible elements:

• Program title: h1.usa-heading-lg
• Description: .usa-prose p:nth-of-type(1)
• Last updated: .usa-tag (contains "Updated" text)

Use AlterLab's selector API to extract these:

import alterlab

client = alterlab.Client("YOUR_API_KEY")
result = client.scrape(
    url="https://www.census.gov/programs-surveys/acs/",
    selectors={
        "title": "h1.usa-heading-lg",
        "description": ".usa-prose p:nth-of-type(1)",
        "updated": ".usa-tag:contains('Updated')"
    }
)
print(result.data)

Structured JSON extraction with Cortex

For typed data extraction without manual selectors, use Cortex. This example extracts key metrics from a Census data table:

import alterlab

client = alterlab.Client("YOUR_API_KEY")
result = client.extract(
    url="https://data.census.gov/table/ACSST1Y2022.S1901?q=median+income",
    schema={
        "type": "object",
        "properties": {
            "median_household_income": {"type": "number"},
            "margin_of_error": {"type": "number"},
            "geography": {"type": "string"},
            "year": {"type": "integer"}
        }
    }
)
print(result.data)  # Typed JSON output: {"median_household_income": 74580, ...}

Cost breakdown

US Census pages typically require T3 (Stealth) due to header validation and light JS challenges. AlterLab auto-escalates tiers—you start at T1 and only pay for the successful tier. See AlterLab pricing for full details.

Note: For most Census pages, expect T3 usage ($0.002/request). Auto-escalation prevents wasted spend—T1/T2 attempts that fail don't incur charges.

Best practices

• Respect census.gov/robots.txt: disallow paths like /api/ but allow /programs-surveys/
• Implement rate limiting: 1 request/second per IP to avoid triggering protections
• Handle dynamic content: AlterLab's T4 tier renders JS-heavy tables; use wait_for_network_idle=true
• Rotate User-Agents: AlterLab does this automatically via proxy pools
• Monitor responses: check for 429/403 and adjust concurrency

Scaling up

For large datasets:

1. Batch requests: Use AlterLab's /batch endpoint for 100-url chunks
2. Schedule recurring scrapes: Cron expressions via AlterLab's Scheduling API
3. Store results: Stream JSON output directly to data warehouses
4. Handle pagination: Census tables often use &page= parameters)