Back to Community
New Features: Forum Tagging, Improved Search, and Interesting Posts

The forums sure look different today, don't they! I'll start with a quick overview of the features, then go into the design philosophy we followed, and finish off with some more nitty-gritty details.

Tagging

By adding tags to posts, we hope to make it easier for you to find content that is interesting and relevant.

Every post in the forums can have up to 5 tags, chosen from a list (currently there are 32 tags). The tags show up on the main forum page and on the individual post page. The post's author can choose (or edit) tags, and the forum moderators (Quantopian employees) can also add or edit tags.

If you click on one of the tags you are taken to a listing of only posts with that tag. If you're interested in pairs trading, for instance, you can easily see all the posts with that tag. You can also search within a single tag.

Improved Search
The search engine is greatly improved. When you're looking for help, an idea, or some reference you once read the search engine will now return posts that are more relevant.

Search changes:

  • New, more powerful search engine on the back end
  • Now indexed on post and replies as one unit, instead of indexing each reply separately
  • You can run a search, click on a result, then click the back button, and your search history will still be there
  • You can search within a specific tag
  • You can search by author

Interesting Posts

As the community has grown, it has become harder to keep track of new, high-quality content. We're trying to improve that by creating the "Interesting" tab. The contents of that tab are curated by Quantopian employees. If we think a post is particularly interesting - it will start showing up in the Interesting tab. The contents of the tab are still newest update to oldest.

The "Newest" tab is uncurated. That's everything, newest update to oldest. The "Newest" tab is the same behavior as the old forum system.

Design Philosopy

The original design of the forums, three years ago, was "keep it simple." Forum software is some of the oldest, most mature technology on the internet, and there are a million different features and variations. We stuck with the bare minimum we needed to start building a community, and we told ourselves that we'd add features as we needed them. We've been pretty happy with the bare minimum and have only made a few minor tweaks so far.

The problem that we're trying to solve is one of those problems that you love to have. There is so much content now, and so much new content each day, that it gets harder for a given visitor to find content that is relevant to them.

Given that problem, the improved search is a pretty straightforward change. You want to find all the posts about Tony Cooper's ideas? Now you can.

But what about less directed consumers? Just regular browsing? We have a lot of people who come to Quantopian to just "see what's up." Those people may not want to see the new person struggling with leverage, or the latest finicky fetch_csv troubleshooting session. We're trying to solve that aspect of the problem with curation in the Interesting tab.

And, what about the in-between? You want to learn more about pair trading, but you don't have a specific question in mind? That's really what tagging is about. We could have gone a step farther and broken the community forums into distinct forums, but that felt too aggressive. We want it to feel like one unified community, not a bunch of splintered conversations. The tags should permit people to browse and search the flavors that are particularly interesting, but not cause the community to split. We also considered making tags entirely free-for-all rather than a set list. The challenge there is that one person thinks of one set of tags, and the second person thinks of a second set of tags, and then no one can find each other's content. By limiting the list of tags we can create a common framework. Clearly, there is no perfect answer, and tradeoffs needed to be made.

It's worth mentioning what we're not trying to solve, at least not yet. We don't have any "reputation" features where individual members can accrue reputation from fellow members. That type of feature is used for driving high-value replies, and we don't think we have a problem with low-value replies. We also don't have a notion of voting up a particular reply, like a Q&A website. We are sticking with our more flowing conversation style. Similarly, we don't have threaded replies, though this one may come. Some of our conversations zig-zag a little, and threading might be a future improvement.

Details

Here is the current tag list, and the general idea behind each one. If you have a tag that you think is needed, comment or send us feedback.

  • API - specific questions or ideas about the Quantopian API
  • asset allocation - discussions or algorithms that cover asset allocation
  • collaboration request - for people looking for teammates in algorithm building
  • contest - posts about the contest, getting into the contest, contest entries, etc.
  • ETFs - discussions or algorithms that involve ETFs
  • feature request - changes you want in the Quantopian platform
  • fetcher - questions and ideas on using fetch_csv()
  • finance news - discussions about topical events, stories, etc.
  • finance theory - discussion of ideas, concepts, or events in finance theory
  • getting started - a new community member trying to get started
  • hedge fund - topics related the the Quantopian hedge fund
  • live trading - questions and discussions for paper trading and real-money trading
  • machine learning - discussions or algorithms that cover machine learning
  • market neutral - discussions or algorithms that use market neutral concepts
  • mean reversion - discussions or algorithms that involve mean reversion
  • momentum - discussions or algorithms that involve momentum
  • pairs trading - discussions or algorithms that involve pairs trading
  • papers - links to interesting academic papers
  • portfolio optimization - discussions or algorithms that cover portfolio optimization
  • Q meta - for discussions about Quantopian itself - business model, contest rules changes, things like that
  • Q updates - to be used by Quantopian employees with news about Quantopian events
  • research - questions, ideas, comments, and discussion of the research platform
  • risk management - discussions or algorithms that cover portfolio optimization
  • seasonal trends - discussions or algorithms that involve seasonal trends
  • seeking help - anytime the author is looking for help. this is going to used a lot! It's often paired with another tag or tags.
  • sentiment - discussions or algorithms that involve sentiment analysis
  • Show Q - to be used anytime you've built something neat you want the rest of the community to check out. It's inspired by Show HN.
  • stat arb - discussions or algorithms that involve stat arb. This is a broad umbrella.
  • statistics - for posts that go into statistical concepts
  • tools and tips - for code snippets, reusable functions, quick advice, and other handy bits that can be reused by other community members
  • using data - a broad category related to ingesting data in your algorithm
  • value - discussions or algorithms that involve a value thesis or more generally using corporate fundamental data
Disclaimer

The material on this website is provided for informational purposes only and does not constitute an offer to sell, a solicitation to buy, or a recommendation or endorsement for any security or strategy, nor does it constitute an offer to provide investment advisory services by Quantopian. In addition, the material offers no opinion with respect to the suitability of any security or specific investment. No information contained herein should be regarded as a suggestion to engage in or refrain from any investment-related course of action as none of Quantopian nor any of its affiliates is undertaking to provide investment advice, act as an adviser to any plan or entity subject to the Employee Retirement Income Security Act of 1974, as amended, individual retirement account or individual retirement annuity, or give advice in a fiduciary capacity with respect to the materials presented herein. If you are an individual retirement or other investor, contact your financial advisor or other fiduciary unrelated to Quantopian about whether any given investment idea, strategy, product or service described herein may be appropriate for your circumstances. All investments involve risk, including loss of principal. Quantopian makes no guarantees as to the accuracy or completeness of the views expressed in the website. The views are subject to change, and may have become unreliable for various reasons, including changes in market conditions or economic circumstances.

6 responses

Nice! looks great and definitely indexing will help leaps ... wondering if there is a quick way to only view posts with algo's attached without scrolling around searching for the algorithm attached logo?

Hi Dan,

Regarding the improved search, I'm guessing there is just a big honkin' database under the hood. So, does the search have an associated search syntax, with boolean support, to access all of the fields? Have you provided direct access to the database, or is there an extra layer that simplifies the search?

For example, how would I search by post author, with an additional keyword filter applied to the post title, and restrict the search to a specific range of dates?

Grant

@Grant Kiehne Thank you very much

good

Grant, we haven't built that rich of a feature set into the search engine. Right now the only "special" search feature is the ability to limit your search to threads with a single specified tag. We're working on making it so you can limit your search to items with two specified tags. Other than that, no boolean/title/author/etc.

Taking a step back, the design goal was to make it so that most searches find highly relevant results. I tested the top couple dozen searches on the old system and the new system, and the new system was giving back much, much better results. (That's a qualitative assessment, not quantitative.) However, I know that it's still not serving 100% of search desires.

@Grant, Google gets you a bit closer to what you want. Try: site:https://www.quantopian.com/posts "something weird"