Aug 01

According to this brief study done in 2006, not very:

Based on the evidence we’ve gathered here, it’s safe to say that no external metric, traffic prediction service or ranking system available on the web today provides any accuracy when compared with real numbers.

I wonder if things have changed two years later?

Jul 31

“01101100011011110111011001100101″ - 732 results

“01110011011010000110100101110100″ - 246 results

“011101000110100001100101″ - 176 results

“011100110110010101111000″ - 106 results

“0110100001100101011011000110110001101111″ - 9 results

“01101100011001010110010101110100″ - 8 results

“0110100001100001011100000111000001111001″ - 8 results

“011001110110111101101111011001110110110001100101″ - 4 results

“01100010011100100110100101110100011011100110010101111001″ - 0 results

Jul 27

Decided to try out one of these book social networks to see if I can find some important books I should be reading. The idea is you enter the books you’ve read and rate them, as well as the books you want to read. Then the site will show you people who like similar books and you can hopefully stumble upon some books that they’ve read that you will like.

The three main sites I found were Shelfari, GoodReads, and LibraryThing. The idea is pretty cool, because a good book recommendation, unlike a movie recommendation, can possibly change your life. So the chance of getting a huge payoff from these services is relatively very high. Not bad for spending no money and just a couple minutes or even an hour or two entering your book list and exploring the shelves of others.

My primary goal is to find business and technical books. Finding good fiction or literature to read is not a problem for me–there’s plenty of it out there. Really what I’m looking for is books that have been read by people I respect in a business/technical role. That’s why I think a service like this could be more valuable than Amazon’s recommendation engine. I’m looking for quality books read by quality people. Amazon generally recommends popular books. The books I want to read don’t necessarily have to be popular among the masses, just popular among the people who I want to learn from. 

I started with Shelfari, then checked out Goodreads, and then on to LibraryThing.

They are all pretty good. I will update this post later with more detailed reviews, but I have to say if you sign up for one you might as well sign up for them all. You will increase your chances of finding “diamond in the rough” books that way. Of course, it takes time to input your books into all 3 of them, but there could be a quicker way to do that(I’m gonna look into existing sites, otherwise might roll one myself).

The traffic stats:

Compete.com shows goodreads pulling away from the comptetion.

Google Trends

shelfari(red), goodreads(blue), librarything(orange)

Google Trends for Websites also shows goodreads in the lead and growing fastest.

Jun 02

A random assortment of cool things I’ve been using the past few weeks:

  • J (a mathematical programming language–free)
  • RankChecker (monitor your websites search rankings for certain keyword–free Firefox extension)
  • BeautifulSoup (python screen scraper)
  • SEO for Firefox (free firefox extension that provides lots of seo goodies)
  • Project Euler (always a fun way to spend some time)
  • Actuarial Outpost (great stuff for those in the actuarial field or planning to be)
May 27

When are they going to put weather data on Google Maps? They’ve got satellite views, terrain views, traffic views; why isn’t there a weather view?

May 16

I think Wigix.com could be a real eBay competitor. On Wigix, you build a “portfolio” of items you own. Then they tell you the latest value of each item in your portfolio. So you can add a ton of items you own like a car, computer, cell phone, mp3 player, watch, etc., and then you can check the value of each item. If the price is right, you can list it for sale. What’s great about Wigix is that you can create your portfolio and then come back 2 months later (when you need cash) and see the prices you could get. The fees are much lower than eBay as well.

I wouldn’t be surprised if Wigix takes off or gets acquired within eighteen months.

May 10

I wish Gmail made it easier to filter Junkmail.

I would say there are three main types of email:

1. Email. Messages you want to always read: emails from coworkers, family members, your favorite mailing lists, etc.
Gmail does a great job with these. You can easily read your mail, star it, archive it, etc.

2. Spam. Messages you don’t want to ever read, and never, ever solicited: Viagra ads, Nigerian business deals, phishing scams, etc.
Gmail does a great job of filtering spam.  I get between 80-100 spam messages a day. I would say less than 1% make it through to my inbox.

3. Junkmail.  Messages you probably never want to read, but may have solicited months or years ago: updates from eBay, Yelp, [any website that you signed up for at one time or another]…
My inbox is absolutely filled with junkmail. Technically it’s my fault, because I’ve signed up for at least a hundred web 2.0 apps in the past year, but why must it be so hard to filter this? Right now I have 7,641 unread emails in my inbox. I would guess that 80% of those are junk mail. They are from sources like the The Funded, ABC’s The Note, Slashdot, Yelp, Deval Patrick, Hillary Clinton, Barack Obama, Barnes & Noble, Buy.com etc. Sometimes I’ll sign up for some list without even realizing it. Other times I’ll actually want a daily recap of some blog emailed to me….for about a week. The end result is an inbox that is so inundated by unread mail that I pretty much just leave everything in my inbox–and never know how many true unread messages I have. This is frustrating.

Gmail offers a quick fix to filtering an email list like these ones. For instance, I can set all messages from “buy.com_offers@enews.buy.com” to be either deleted or archived automatically. The problem with this approach however, is that I would have to create hundreds of these filters, and keep creating new ones each month. This wastes a lot of time.

Another solution is to mark these messages as Spam. I am wary of doing this. I believe Gmail’s spam filter works via a Bayesian Spam Filter that calculates the probability of a message being Spam by comparing its attributes to previous messages marked as Spam. I think that Gmail aggregates the Spam reporting of its users to filter spam for each individual. If this is the case, than by me marking a message from SocialPicks as Spam, I would increase the likelihood that emails from SocialPicks to other users–some of whom may actually want to read messages from SocialPicks–would also be marked as Spam. I don’t want to do this, because generally I signed up for the email list in the first place and its not SocialPicks’ fault that I actually don’t want to be signed up for it anymore. (If anyone knows the details of how Gmail’s spam filter actually works, and whether or not me marking these types of emails as Spam would actually cause these negative effects, please let me know.)

So if it takes too long to create filters for every junk mail feature, and if I don’t want to mark messages as Spam, what are other possible solutions?

Mark messages as “Junk Mail”. Junkmail is neither email I want to read but also not Spam. Not only could commerical messages from companies be marked as Junkmail, but any type of email list: say a fraternity email list that you don’t care to read after being out of college for a year. It would automatically send these messages to your archive using the same techniques of the spam filter. This feature could also be called “Autofilter”, I don’t really care.

An “Unjoin List” button. Most of these Junk Mail messages have a link to stop receiving these emails. But these links are usually a pain in the a** and require visiting a web page and filling out a form. If I’m not going to take the time to read your email, I certainly don’t want to take the time to visit your website and fill out a form telling you why I don’t want to read your email.

Personally, I would like both of these features to be implemented. Sometimes I want to unjoin a list, sometimes I’d rather not leave a list, but don’t want to create a filter to autoarchive it. I would want the latter in situations where I don’t want to offend the creator of the list or I want to have a copy of the list messages in the unlikely event that I would ever need to reference them.

Filtering Spam has been conquered. Can we conquer Junkmail next?