hey chatgpt is it possible to determine the number of appraisers in us by removing duplicate licenses, thus beginning with the federal ASC appraiser list, then cross checking for duplications among state appraiser lists?
Yes. In principle, that's one of the best ways to estimate the number of individual appraisers rather than the number of licenses. The key is understanding what the data already contains.
The Appraisal Subcommittee (ASC) maintains the National Registry of Appraisers, which is compiled from state licensing agencies. An individual licensed in multiple states generally appears multiple times because each state credential is a separate registry record.
A practical deduplication workflow would look like this:
These are probably the same person.
But:
These could be two different people unless you have additional identifying information.
Yes. In principle, that's one of the best ways to estimate the number of individual appraisers rather than the number of licenses. The key is understanding what the data already contains.
The Appraisal Subcommittee (ASC) maintains the National Registry of Appraisers, which is compiled from state licensing agencies. An individual licensed in multiple states generally appears multiple times because each state credential is a separate registry record.
A practical deduplication workflow would look like this:
- Start with the ASC National Registry.
- Download the full registry (if available through their data access).
- Each row typically represents a state credential, not necessarily a unique person.
- Identify duplicate individuals.
- Group records using combinations such as:
- First name
- Middle initial
- Last name
- Credential number
- State
- Business address
- City
- Exact-name matches are a starting point, but you'll also need fuzzy matching for cases like:
- Robert Smith vs. Bob Smith
- John A. Doe vs. John Doe
- Address changes
- Group records using combinations such as:
- Count unique people.
- Once duplicate licenses belonging to the same individual are linked, count distinct individuals instead of records.
- Validate against state rosters.
- Some state rosters contain additional identifiers or more current information than the federal registry.
- They can help resolve uncertain matches.
Challenges
The biggest challenge isn't the counting—it's entity resolution (determining whether two records belong to the same person). For example:| State | Name | License |
|---|---|---|
| Texas | Jane M. Doe | TX12345 |
| Oklahoma | Jane Marie Doe | OK67890 |
But:
| State | Name | License |
|---|---|---|
| Florida | Michael Johnson | FL11111 |
| Georgia | Michael Johnson | GA22222 |
Accuracy
Depending on the fields available, a careful matching process can often achieve well over 95% accuracy. If you have access to stable identifiers (such as a national registry ID or another persistent identifier), accuracy can approach 100%. Without them, you'll typically use probabilistic record linkage.Is it worth using state lists?
If your goal is simply estimating the number of unique licensed appraisers nationwide, you may not need to scrape all 50 state lists. Since the ASC registry is built from the states, it is already the primary consolidated source. State lists are most useful for:- verifying ambiguous matches,
- filling in missing information,
- identifying records that haven't yet propagated to the federal registry.
pandas, recordlinkage, and rapidfuzz) that would automatically identify likely duplicate appraisers across all states and estimate the true number of unique licensed appraisers in the U.S.
