I’m storyboarding a concept for building search criterion in a natural way. I did this in PowerPoint so it’s not exactly how I want it to look, but I think the GIF gives a good general idea of how I want it to look.
3 Likes
This is a cool idea. We use Anvil to let users search for items using various kinds of pre-assigned filters, and I could see this kind of guided semantic search feeling intuitive. Have you taken a crack at making it work in Anvil?
Thanks 
What kind of company do you run?
I haven’t stared building it yet, though I plan on doing it soon! I’ll be sure to post a link to the app when it’s in a semi working state.
We organize and clean product catalog data and all the associated options. A key part of this involves a whole lot of normalized labeling by way of AI (natural language and image analysis) and human review & refinement.
We use the clean data in a few ways, but imo the most interesting is our main Anvil app. With the app, users in-store can explore product catalogs in a “human-intuitive” way by exploring and layering labels. These are implemented as buttons in Anvil, and users can work their way in to filter down to the products that check all their boxes. Beats the hell out of flipping through a dozen 200+ page printed catalogs.
It’d be interesting to help users build a one-sentence description of what they’re looking for, and our labeling data’s well-suited to it. Something like “show me [chairs] that have [arms] and are made of [wood] or [aluminum].”
What kind of company are you running? How would you use a search feature like this?
That is a pretty cool solution, though I’d argue my methodology is more intuitive and visually appealing.
Completely objective and non biased opinion