API basics
HMDA is a GET API that lives at https://api.consumerfinance.gov/data/hmda.
- You can query an entire dataset.
- You can query dataset slices, which are pre-loaded views we think are interesting.
- You can query concepts, which describe types of information found in the dataset.
Datasets
The endpoint for querying all data begins with /data/.
| Endpoint | What it does |
|---|---|
/data/{dataset-name} |
Gives all information about a dataset |
/data/hmda |
Gives mortgage data from 2007-2012. |
Concepts
Concepts are analogous to variables, or column headers in a spreadsheet. Concepts have properties, which describe all the possible values. You can also specify concepts in any of the supported file formats like so: /data/{dataset-name}/{contept-name.extension}.
Slices
Think of slices as tables in a relational database. Every dataset has many slices representing different views of it, which you can use to construct advanced queries. The endpoint for
every slice is /data/{dataset-name}/{slice-name}.
You can even request a slice in HTML, XML, JSON, JSONP, or CSV. Just append the filename extension to the endpoint like so:
/data/{dataset-name}/{slice-name.extension}.
Putting it all together
Every year, the Federal Reserve finds interesting trends in HMDA data and publishes them in a report. Let’s say you want to replicate some of their 2012 highlights in JSON.
To compare refinances and home purchases in 2012, you would send the following query to the API:
https://api.consumerfinance.gov/data/hmda/slice/hmda_lar.json?#!/property_type=1,2&action_taken=1&select=as_of_year,loan_purpose_name,count§ion=summary
To see changes in the FHA loan market for 2012, you would send the following query to the API:
https://api.consumerfinance.gov/data/hmda/slice/hmda_lar.json?#!/lien_status=1&loan_purpose=1&action_taken=1&select=as_of_year,loan_type_name,count§ion=summary