Steemfilter is my project aimed to help Steemit users find new quality posts. Today I am happy to announce that a brand new Steemfilter’s version is rolled out.
Steemfilter’s goal is to make potential quality posts shine as bright as they deserve. To attain it, it filters out the posts that require less effort to produce, like short posts with no images or copypasted posts, and presents the remaining posts more prominently for the reader's enjoyment.
New features and improvements
- Full redesign using Twenty Fifteen theme which turned out to be the most convenient by design and simplicity. This native WordPress theme is stable and well-supported.
- Drastically improved load speed, as promised.
- Added controls allowing to change some of the filtering parameters.
- There is now a checkbox allowing to load only posts tagged #introduceyourself. I dedicate some time to curating new members' posts in this category and this addition makes it so much easier.
- Google language detection is replaced with PHP Text_LanguageDetect library to improve speed and reduce costs. This library works less precisely than Google’s and it turned out that it doesn’t support Chinese and Korean, so I will search for better solutions in the next version.
- Now the default post age is 1 minute with the possibility to change it to 5 or 10 minutes (which slows down the script a bit but is still acceptable).
- Project blog and contact section are added.
If you want to get maximum curation rewards from voting I recommend Vote in X Minutes Bookmarklet by @lukestokes.
The stats say Steemfilter is used regularly and I am sure this rate will grow substantially with the improved speed.
How it works
Currently Steemfilter uses the following filters:
- Language detection (so far English only, later I’ll add more languages).
- Short post check. It’s so much easier to produce short posts and it could look more profitable, but my experience says a short post gets a good payout very rarely, at least for new Seemit authors.
- Post with no images. The same reason as above — finding a good licensed image requires effort.
- Plagiarism check. So far the script just checks if a post wasn’t marked by @cheetah bot.
The code is written in PHP and uses Steem API by @goodkarma. Steemfilter is built on WordPress platform which allows easily adding website-related features.
Future plans
- Add tags
- Support more languages
- Better plagiarism check
If the project’s user base will continue growing, I’ll experiment with text analysis tools like readability tests, topic detection or English grammar check.
I invite you to use Steemfilter to support new authors and find interesting posts to vote for. Share your experience in comments or contact me directly. If you like what I do, please support me by voting for this post. Resteem it to help get more feedback!
Credits
Nice to meet you dear freind, welcome to steemit, hope you are comfortable here,This is a new home for lovers
of bloggers and the best platform today,Very good intro,I like that,No forget to Visit my profile For Info and tricks about cyrptotrading.
This concept is really cool. Anything that will make the blogging experience more productive is always welcomed.
This may help minnows get more attention; and, possibly will be something that can be tweaked to include the upcoming Smart Token.
Thanks for sharing!
Peace.
Thank you for the feedback!
✌
This looks awesome and the interface is beautiful from the thumbnail! It's all the developers that put in time to build on top of this block chain that make Steel it the awesome place that it is. Be sure to support the developers everyone, they are truly what makes this place continue to be great.
That is amazing! I was already getting fed up browsing all the effortless content on 'new' tab.
That's why I've started it, the new tab looks awful :)
useful thank you
I just found one, without that tool :) keep up the good work, following for more :)
Thank you!
Very post 👏
Very cool, this is going to be useful! Thank you for your dedication and hard work :)
Thank you!
This is a great idea thanks for your work on this will check it out for sure. Be well
Let me know how it works for you :)
i try using it but... doesn't work as advertised
Thanks, what exactly ?
That is so brilliant. I am very grateful for that! Love it.
Very informational post for me
You have an awesome work done
check my new post
https://steemit.com/uikit/@abkktk/resource-ui-ux-tool-for-web-service
Such comments are considered spam here.
@steemflagrewards
I hope that we will also get success with your help my dear friend
Is there any way a filter could exclude posts that have been upvoted by specific accounts. I would love to eliminate the posts being upvoted by the bots and concentrate only on my own feed (those writers I have selected to follow). I would also like to search and find new people to follow without the hassle of going through their content to check for bots.
In your feed there are only those authors you have selected to follow.
In addition, there are many dozens of bots and no way to detect one except a list compiled and supported manually. Also, many good posts may have been voted for by bots for various reasons. And there are some good bots too.
Thank you for your reply. If I could compile my own 'exclusion list' containing bots which I deemed to be the type which upvotes inferior content, that would work (and I think it would be the same for many people out there). Furthermore, good content like yours for instance, is not upvoted multiple times by several of these bots. In my own opinion, any bot which upvotes someone's post without a human intervention, is a bad type of bot, and I would rather eliminate all content providers who use such devices, as a metter of principle. In short, an upvote should reflect the decision of another human being, and not be the action paid for by the person providing the content.
If more users will be asking for this feature, I may implement it. It could be interesting to see only posts unsupported by bots and vote for them to help. But now I am thinking about text analysis to improve the algorithm of selecting quality posts.
I ndeed, it would be interesting too, to sample content which is unsupported and that which is supported by 'defined blacklisted' bots, to discover if there is significant quality discrepency too. The news about your plans for text analysis, is very exciting. No doubt very difficult to implement too. Thanks for your thoughts and hard work.
I was commenting on a thread link-1, discussion about feature in shadowbot and really considered letting you know about the @shadowbot team work - their interface might do well with a touch of your filter as I mentioned in my comment under that thread.
Then a few hours later, my comment was upvoted by @thing-2 a 'relative' of @gentlebot
Both these bots upvote only comments which have passed some sort of algorithm checking for 'quality' - the process is discussed in summary at link 2. It apppears to be an implementation of a sort of AI - an algorithm like the one you are planning. If there is a chance, perhaps you can look into my two links.