The Scenario
You are an analyst at a residential development firm. Your supply-demand model needs 36 months of Census construction data: new single-family housing starts, multi-family starts, completions, and permits. The model is overdue and your manager asked for a draft by end of this week.
That data lives in the Census Economic Indicators Time Series (EITS), updated monthly. You have pulled it once before, last year, by downloading the EITS API output to a CSV. The CSV had 40 columns and you spent an afternoon finding the four you needed.
The bad version:
- You go back to the Census EITS API documentation
- You find that the endpoint has changed slightly since last year
- You reconstruct the query, download the bulk file, and filter to the four series you need
- The dates are in a format your model does not expect (YYYY-MM format vs. YYYYMM)
- You run a find-and-replace and break the date column in three rows
- It is Friday. Your manager wants the draft. You have a CSV with a broken date column and no model.
The fast version is one prompt.
The Easy Way: One Prompt in SheetXAI
SheetXAI is an AI agent inside your spreadsheet that knows the EITS dataset, its series identifiers, and its date format conventions.
Open the SheetXAI sidebar and type:
Fetch Census EITS residential construction data for the past 36 months — single-family starts, multi-family starts, completions, and permits — and write each month as a row with seasonally-adjusted values into my sheet. Use YYYY-MM-DD date format in column A.
SheetXAI calls the EITS API, identifies the right series codes, formats the dates consistently, and writes one row per month into the sheet. The four construction series land in their own labeled columns.
What You Get
A clean monthly time series, 36 rows long:
- Column A — date in YYYY-MM-DD format
- Column B — single-family starts (seasonally adjusted)
- Column C — multi-family starts (seasonally adjusted)
- Column D — completions (seasonally adjusted)
- Column E — permits (seasonally adjusted)
These are the official Census Bureau New Residential Construction release numbers, the same figures cited in housing market research and Federal Reserve reports. Your model footnotes write themselves.
What If the Data Is Not Quite Ready
Time series from government APIs have predictable edge cases that can break a model if not handled upfront.
When you need seasonally adjusted and not-seasonally-adjusted side by side
Your econometrician wants both versions for their X-13 decomposition.
Fetch Census EITS single-family starts for the past 36 months in both seasonally-adjusted and not-seasonally-adjusted form. Write the date in column A, seasonally-adjusted values in column B, and not-seasonally-adjusted values in column C.
When you need to extend the series back further than 36 months
Your model needs 60 months to capture a full cycle.
Fetch Census EITS single-family starts, multi-family starts, completions, and permits for the past 60 months, seasonally adjusted, with one row per month. Write into columns A through E.
When you need retail sales data instead of construction
Your firm is shifting focus to retail real estate and wants the EITS monthly advance retail sales series instead.
Pull Census EITS monthly advance retail sales for the past 24 months and write the series name, period, and value into columns A, B, and C of my sheet.
When you need to join construction data with interest rate data from another source for the model
Your supply-demand model needs mortgage rate data alongside starts to build the rate-sensitivity curve. You have 36 months of mortgage rates in the Rates tab.
Fetch Census EITS single-family starts and completions for the past 36 months, seasonally adjusted, and write date, starts, and completions into columns A, B, and C. Look up the 30-year fixed mortgage rate for each month from the Rates tab (date in column A, rate in column B) and write the rate into column D. Flag any months where the mortgage rate is missing in column E.
The pattern: official construction series, your own rate data, and the joined model input in one prompt.
Try It
Get the 7-day free trial of SheetXAI and open any sheet where you need Census economic time-series data, then ask it to pull the EITS series you need by plain-English description. The Census Bureau integration is included in every SheetXAI plan. For related workflows, see how to pull County Business Patterns by NAICS code or the Census Bureau in Google Sheets overview.