Our geothermal savings estimation API, LoopLink GSE, recently received an update to improve accuracy and performance. We will start with what got better then talk a little about technical changes that will impact customer implementations.
More Accurate Utility Rates
In version 1 of the GSE API, we took statewide average utility rates and applied them to any search which didn’t include custom rates. In this update, we now search (in the US only) for local electric rates and apply those to the estimate. So, if your customer searches their home address we can typically identify the correct electric utility provider and rate.
We have yet to find a house level rate lookup tool for natural gas, fuel oil or propane. So those energy prices are still a statewide or regional average depending on data availability.
We know it isn’t optimal but due to the difficulty in finding consistent information for all countries, searches outside the US will resolve to the US national average for a given energy type. As demand increases in other countries we will look to improving our defaults. For now, know that if you intend to support other countries your best results will come from asking the homeowner for utility pricing.
Know the Place?
A place object has been added to the response which enables you to identify the location we found for the estimate. At first blush this doesn’t seem that valuable. After all, the user provided the location to initiate the request. The value is that now you have a way to provide feedback to the user to ensure that the place we found was in fact the place they searched.
Technical Note: We moved geocoding attribution data into this object… it just made logical sense to tie the attribution to the result.
Updated Load Model
We are always looking at our load model and working to provide more accurate results with the limited information we require from homeowners. In this update we did more work on tweaking our advanced parameters to account for duct placement, seal and insulation in a more accurate way.
Better Security and Accounting
Version 1 was a little naive in its implementation of Cross Origin Request Sharing (CORS). We have bolstered this method and improved our ability to identify sites that are registered to make requests as well as improved the accuracy of our per site request counts. This will greatly simplify your ability to make requests against our system and allow us to provide you with more accurate usage statistics.
Logging and Retrieval
We noticed that an awful lot of homeowners come back and enter the exact same information into the system multiple times. Rather than rerun these requests we have implemented a logging function that will store requests for up to one month.
We provide a unique estimate ID with each response that can be requested directly. So, now you can get a homeowner’s email and send them a link back to their estimated savings or send a local dealer a copy of the link with the homeowner estimate.
We do not accept or store contact information so it is up to your team to fill in those blanks.
GET or POST
In LoopLink GSE 1.0 we required all requests to be sent as a POST. This is no longer the case. You may now submit your data with a GET or POST. All other request methods are restricted.
Change Your Request Structure
In our first iteration we tried to build GSE to be as permissive as possible. We allowed users to just send us a location field that contained anything from a complete address to just a postal code. The tool worked but the search was buggy and while it always returned a result, the result may have been for the wrong place entirely.
To deal with this issue, we now ask for structured requests so we can more efficiently and precisely search locations. So what was previously sent to us as:
- location=1234 Any Street, My Town, SD 12345 USA
Is now sent as:
- street = 1234 Any Street
- city = My Town
- state = SD
- postal_code = 12345
- country = USA
You don’t necessarily need to send us every piece of information but we we now require that you include the country in a search based strictly on postal code.
Search Coordinates Directly
In the ideal interface, your user will run a location search in the browser through Google Maps or a service like Zillow (which provides some square footage information). Which means there isn’t a good reason for us to geocode the location on our end. You can send us coordinates directly which does speed up many of our results. Just remember that any search containing ‘lat’ and ‘lon’ fields will be reverse geocoded to establish rates and other location information.
In the update, you no longer need to provide the ‘geocoded’ flag. We automatically detect the 'lat' and 'lon' fields and handle the request accordingly.