Thread by @Neeva

1/ Have you heard? Bias Buster is now available in #Neeva's main search tab!

🔎 Try a search here: neeva.com/search?q=calif…

And if you're wondering how we crawled and evaluated topics to create our 5 point scale slider, stay tuned! 🤓

We dive into it on this thread 🧵…

2/ Our goal? 👉 Show a variety of POVs on particular news topics.

To reach this goal, we categorized results based on 5 buckets to ensure a smooth experience while interacting with the slider. This includes:

🪣 Far Left
🪣 Left Leaning
🪣 Center
🪣 Right Leaning
🪣 Far Right

3/ So, how do we categorize our results to fit these buckets?

By using third party media bias tools, such as @AllSidesNow and @MBFC_News.

Each result is categorized by its respective domain.

4/ We use these ratings to filter and re-rank what results show in what bucket.

Categorizing domains is not a perfect science, so we make sure to show results at most one bucket away from the selection on the slider.

5/ With this bucketing implementation, we need to ensure that we have domains to serve from all of these perspectives.

We collected a variety of domains from each of the 5 buckets, and pulled the respective sitemaps.

These sitemaps are fed into our crawl pipeline.

6/ Previously, crawling and indexing one URL into Neeva’s own index took more than 2 weeks after the URL’s discovery.

Apparently, this is too stale in terms of serving news pages.

7/ In order to serve news pages in a fresh way, we build our fresh crawl-indexing pipeline.

Every hour, we crawl and index URLs from a couple of sources, including:
📌 Sitemaps
📌 Twitter feeds
📌 API crawl, etc.

From there we fast-track these pages into our Koala indexing. 🐨

8/ To utilize Bias Buster to its full potential, we implemented triggering logic.

This allows the slider to show if there are a variety of results to view on the spectrum.

We determined these queries are typically ones that have a high news intent, as well as political intent.

9/ So, we:

1️⃣ Probe the result sets pulled from the buckets to gain intuition on variety
2️⃣ Check political intent & topicality
3️⃣ Check if the query has any identified intents for which we shouldn't trigger on

10/ Here's an example...

If the query includes a site restrict, we wouldn't want to display Bias Buster, since the ultimate intent is to see results from that site.

11/ Overall, Bias Buster gives an opportunity for our US users to explore different perspectives on the political spectrum when available.

Head over to neeva.com to try it out and let us know what you think!