| Overall: | 17 | 4 |
| Design: | 10 | 0 |
| Navigation: | ||
| Usefulness: | 8 | 0 |
17
4
What is it?
TweetBlocker is the latest Twitter Ruby on Rails application from Hashrocket Labs. Tweetblocker is a free resource for Twitter users and application developers. Using highly–sophisticated Bayesian filtering, we catalog and rank the top spammers on Twitter, allowing users to quickly and easily figure out who’s spamming.
Mission
Our core mission is to clean up the Twittersphere. However, we cannot do it alone, so we've built a system that both users and application developers can proactively utilize. Users can use our bookmarklet tool to report any spammers as they navigate through their experience on Twitter. This gives people the peace of mind knowing that we're cataloging their submissions and taking care of the spam for them.
How to use it?
TweetBlocker is very easy to use. There are 4 different ways to use TweetBlocker:
1. Check spammer from homepage
2. Login and automatically generate a spam report
3. API for developers. So you can both retrieve grades and report spammers.
4. Bookmarklet tool. Report spammers with a click of a button as you intuitively navigate through Twitter each day.
Check spammers from homepage. With this tool users can instantly retrieve a grade for any person they think might be a spammer.
Users can login and quickly retrieve a spam report that we generate on-the-fly. This report gives the user a holistic view at both the people their following and people that are following them. Users can sort by grade and block spammers permanently in this view. Users can easily report spammers in this view by clicking the button report spammers in the header. This submission adds the spam 'suspect' to our index where run a variety of checks before deeming the suspect as a real spammer.
Web developers and third party application developers can utilize our RESTFUL API to do the following:
- Grade Method (able to retrieve a letter grade for any user)
- Report Spam Method (able to report spammer suspect to our index)
- Rate Limit (currently we limit developers to 100 calls per hour. (Please contact me at jonathan@hashrocket.com if you'd like to increase your API usage.
Users can utilize our easy to use bookmarklet tool. The user can drag and drop the bookmarklet from our homepage or dashboard page into their browser toolbar. When a user comes across a suspected spammer all they need to do is click the Report Spammer button in their browser toolbar. The users submission is instantly added to our suspects index where we perform a variety of checks before deeming the suspect as an official spammer.
- what else do you want this app to do? please leave on our feedback forum at http://tweetblocker.uservoice.com.
- usability of the site. how does it feel to you? is the messaging clear? did you understand the meaning of our app in less than 10 seconds of coming to our site?
- what are some other ways you would like to interact with our app? what would make your life even easier when removing Twitter spam?
- are there any other API opportunities that we've missed?
- have you tried our 'Report Spam' bookmarklet? did it makes sense? did you know how to work the bookmarklet? were our instructions clear enough?
- what are some other ways we can help spread the word on Twitter about our new anti-spam algorithm?
- anything else that's on your mind.
thanks!!!
One last thing that would probably be solved by a FAQ. When I click details to find out about a user I get a nice popup with some relatively (to me) useless information. For example I just clicked on a D graded user and it says:
# This is a fairly new user
# This user does not have a balanced ratio of friends and followers
# This is an active user
but my F users said the same things. What is different about this D vs. an F?
Oops, just noticed a problem too. Some of the pages take so long to load all the results that it's easy to click on the Details link before the page finishes loading. If you do that, the JS has not loaded yet so instead of a nice light/facebox you get taken to the underlying route (e.g. http://tweetblocker.com/candidates/[user]).
Your site looks amazing...love the graphics, color palette, organization, info, simplicity...it definitely looks like it was made "with love". I understood the meaning of the site within the first 10 seconds, but that might be because I read part of the writeup from launchly.
I'm not a REST expert, but looking at your APIs left me a bit uncertain. I want to know what methods those APIs implement, what JSON keys to expect back and what values are acceptable. Perhaps some additional info might be helpful for the non REST experts.
I blocked two people successfully, but there was one user I couldn't block and wasn't really given a good error message as to why it failed. BTW - minor nit, I think your sentence in the error message is incorrect, should say "...people selected" not "...please select"
The 3 steps in "Get started today" are really helpful, except that I have to logout to see them...couldn't find it once I login.
So I tried reporting the user that I couldn't block earlier and after I tried to login to TB, I got a JSON string back:
{
status: "User has been reported as a spammer"
}
Guess now I know what to expect back on some of your REST APIs :-)
Anyway, I login to TB and that one blocked user is still showing up in my spam report but I still can't block them from the UI. So does reporting them as spammers not block them from my account?
A suggestion for your site would be if you could somehow automatically clean-up any users that have been highly reported as spammers from my account.
I think your instructions on the bookmaklet were OK, but I wasn't sure if I was supposed to actually click on the spammer and then click on the bookmarklet or if there was another way...just a bit confused on the sequence there.
How to help spread the word is the million dollar question for any startup, so as soon as I find the answer to that I'll be sure to patent it :-)
A thought might be to create a community that focuses on eliminating spammers from twitter, that might help spread the word. Is your plan to keep the service free? I think that also helps in building a community.
Overall, the site feels clear and simple to use.
Thanks Brian. We're currently using a variety of signal based filters that help calculate our overall grade for a user. We're also continuing advancement on our algorithm to make it smarter and faster. Because we rely so heavily upon data that we get from real tweeters the more people we get using the app the better the algorithm will become.
So funny that you mentioned Tweet Blocker and TwitterMass integration because we currently have plans to integrate the two in the near future. For example, a TwitterMass user can determine a minimum threshold for followers (let's say a B grade) and then we simply audit each new follower to match a B grade or higher for that instance. This helps validate the relevancy and trustworthiness of a new follower right from the beginning. We also get to help clean up the Twittersphere while making a great product that helps people connect on Twitter.
I'll be adding an FAQ per your recommendation above.
We currently load 100 people per page. I'm going to trim that down to 50 people and see if that helps. We're also going to add a button that does something like "Click this button to block all people with this grade". This would alleviate people from having to paginate through 'x' number of pages.
First thoughts: Very cool looking site. I love the homepage.
Confusion: It was not until I started playing with a second account on your site before I realized you where checking who is following me, not who I follow. I would be far more interested in knowing if I accidentally let a spam bot onto my list than if a spammer is watching me.
Blocked Users: You flagged two users with a grade of F on one of my accounts. I checked both of these users on twitter to verify your finding only to discover I was already blocking one of the users. I would not expect your service to show me blocked users.
Rating System: I am having trouble trusting your grading system. One of my accounts is pretty much an auto-follow account with tons of spammers following it. Your system gave most of these users pretty high marks.
It's a fun site.
Good luck!
TweetBlocker is here to save the day! Ok first I must say that I love the homepage. The superhero is awesome and I love the whole "saving the twitterverse" bit.
It was clear what the site did in under 10 seconds BUT I think you may have a problem with some of your verbage depending on your intended audience. It seems like the API stuff is good for developers but the rest of the app is more generally useful to a larger audience but when you say things like "we use sophisticated Bayesian filtering"... well I think you just lost those people. You might want to consider downplaying the techno-babble on the main page a bit and just put a link to an FAQ page that gives more details on HOW the system works for people that want to know (like developers).
That's something I'm really interested in in fact. How does the system go about grading users? I ask because you gave my @launchly account a D!!! The funniest part is I'm guessing at least some of my poor grade is attributable to using TwitterMass to build up followers haha! You seem to have some opposing forces in those two apps!
I'm guessing you already have this planned but I would LOVE to see some integration between TwitterMass and TweetBlocker. I'd like to be able to specify a spam grade threshold (like >= C) and have TwitterMass use TweetBlocker to only follow users that match.