A tool by Firecrawl · uses 8 tools · Government

Permit turnaround peer benchmark

Use when a city building official or permitting director must show council how long permits take versus peer cities. Finds building-permit datasets on Socrata and DataSF, computes issue times by permit type from the raw records, normalises by population and housing growth from Census data, and drafts a benchmark with method notes.

Open in ChatGPT Open in Claude

Save a copy for your team

Tools it calls

Each tool's price and track record across every agent today. Empty means the call worked and found nothing.

ToolPriceMedianSuccessEmpty

Socrata open-data portalsSearch datasetsPrice5 creditsMedian160 msSuccess97.4%Empty7.9%
DataSF open data portalDatasetsPrice5 creditsMedian–Success–Empty–
Socrata open-data portalsDescribe datasetPrice5 creditsMedian–Success–Empty–
DataSF open data portalDatasetPrice5 creditsMedian–Success–Empty–
Socrata open-data portalsQuery datasetPrice5 creditsMedian717 msSuccess90.3%Empty12.3%
DataSF open data portalRowsPrice5 creditsMedian–Success–Empty–
DataSF open data portalRow countPrice5 creditsMedian–Success–Empty–
US Census BureauDataPrice5 creditsMedian617 msSuccess97.7%Empty3.8%

Instructions

What your agent follows when it runs this tool. Paste them into any assistant connected to Firecrawl.

# Permit turnaround peer benchmark

For one city and three to five peers that publish permit records. Turnaround claims are only fair when computed the same way from raw rows.

## 1. Find comparable datasets

1. Search @[Socrata · Search datasets](alexandria:socrata-com/open-data/search_datasets) (query "building permits") for each peer's portal domain. Keep **dataset_id** and **portal**.
2. For San Francisco, find the permits table with @[DataSF · Datasets](alexandria:data-sf-gov/open-data/datasets) (query "building permits").
3. Read field names with @[Socrata · Describe dataset](alexandria:socrata-com/open-data/describe_dataset) and @[DataSF · Dataset](alexandria:data-sf-gov/open-data/dataset). Identify application date, issue date, permit type and valuation columns; never guess them.

## 2. Compute turnaround

4. For each portal, pull the last 12 months of issued permits with @[Socrata · Query dataset](alexandria:socrata-com/open-data/query_dataset) (select application and issue dates, permit type; where issue date in window). For San Francisco use @[DataSF · Rows](alexandria:data-sf-gov/open-data/rows), sized first with @[DataSF · Row count](alexandria:data-sf-gov/open-data/row_count).
5. Compute median and 90th percentile days from application to issue for new residential, alterations and commercial.

## 3. Normalise

6. Pull population and housing units for each city with @[US Census Bureau · Data](alexandria:census-gov/census/data) (acs/acs1 or acs5) to express permits per 1,000 units.

## Write the benchmark

- **Table:** city, permit type, permits issued, median days, P90 days, permits per 1,000 units.
- **Method notes:** fields used per city, excluded records, any city whose dates are not comparable.
- **Where we stand:** rank and gap to the best peer for each type.
- **Sources:** dataset ids, query filters, ACS tables. Say which peers publish no usable dates and what could not be found.

All Firecrawl tools