408 reads

Instant Search and Related Stories: Algolia + HackerNoon

by RichardJohnnMarch 9th, 2022

Too Long; Didn't Read

Now, as soon as you start typing, you will immediately see results show in a large pop-up. The content that can be searched as also been expanded to cover not only stories, but also our tags, companies and the beautiful people of HackerNoon.

Company Mentioned

featured image - Instant Search and Related Stories: Algolia + HackerNoon

You may have noticed this already, but HackerNoon has a new and improved search experience, thanks to Algolia. Lets break down this this much improved search and story curation experience!

Now, as soon as you start typing, you will immediately see results show in a large window. Character by character they evolve. The content that can be searched has expanded to cover not only stories, but also our tags, companies, and the beautiful people of HackerNoon. ✨

Previously we noticed that a sizable number of search queries were coming back with zero results - primarily for company names. Now you’ll see them listed in the instant search window, as well as on https://hackernoon.com/search. You will find these new sections if you scroll down past the stories.

Both of these designs are a work in progress. We’ll be adding tabs to at least the /search page, to better communicate that we support searching these categories now and to help you filter out down to the most relevant result. I am also interested in including more text in each record so that searches can match content by say, a story’s text content itself. Algolia offers the ability to show content with matching phrases highlighted and I think this would pair great with searching story content.

Besides search, Algolia also offers (currently two) AI models that can be trained in such a way as to recommend related stories. The two models are intended for e-commerce sites, which names ‘Related Products’ and ‘Frequently Bought Together’, but the ‘conversion event’ in our case happens when a user completes a story. When a logged in user reaches the bottom of a story we log this event to further train the model. This will allow us to better curate which stories the community thinks deserves more readership, as well as, provide personalized reading suggestions based on your reading history.

If you don’t want your reading history tracked to have personalized recommendations, you can opt out at https://app.hackernoon.com/profile/settings.

Stayed tuned as we soon be deploying more types of pages which will also be searchable, like cryptocurrencies with price charts and related stories, both on HackerNoon and around the web. I also look forward to tweaking the ranking algorithm to highlight our best content, as demonstrated by user engagement. This will serve as a testing ground for personalized content and community-ranked stories, as opposed to simply showing the latest stories per tag and the editor-highlighted stories on the homepage.