Shotbeak.com

When I was a child, my mother used to (and still does) make the yummiest soup. I also quite enjoyed dipping some bread into the soup: so much so, that on occasion I would play around and dip two-three slices into it, effectively rendering the soup useless. Too much bread.

Omnomnomnom

Did I expect the soup to turn a congealed wheat? Not at first of course. With my young mind, I knew I liked bread. I liked soup. I liked bread and soup and stuffed some more bread into the soup. That’s how I am starting to feel about Facebook. I keep adding friends; friends keep adding me. And it is starting to fill up, becoming noisy and bloated. Why is that happening? Dunbar’s Number, my friend. Dunbar’s Number.

Robin Dunbar's book. Read it.

Dunbar’s number (proposed by Robin Dunbar) is a theoretical cognitive limit to the number of people with whom one can maintain stable social relationships. What is that number? 150. It has surprisingly shown up in a lot of real-world examples (from nomadic tribes to modern day factory floors). I was wondering if this would apply to online-relationships. Considering that Facebook is about ‘connecting with people you care about’, the data reveals a surprisingly similar trend. Facebook’s stats show that the average amount of Facebook friends are… drum roooll. 130.

Now, I’ve spoken about this in the past. Today, it is becoming more apparent as Facebook’s active users in the USA has dwindled by 6 million users. Maybe it is just a seasonal trend, a momentary dip, or maybe it is indicative of a greater problem? Facebook doesn’t know what to do about this increasing bloat.

People change. You aren’t that great of friends with some of your highschool buddies. You are however, adding more and more people on Facebook, without actually dumping some of them off. Facebook has only been around for the majority of its users for 2-3 years.

Now move forward a few years, and people will have changed. They will have new people who are more important in their lives. This means, Facebook’s average friends count will keep growing.

How has Facebook battled with this? They have recently (and silently) slipped in controls for the news feed that show only people you most interact with. The odd thing is, it doesn’t run across all devices that way! If you have Facebook for Android, you will notice a lot more updates that doesn’t appear in your web feed (m.facebook.com is the same as the web feed). Does it work. Not really. It sucks in a certain way to determine what I deem important. Just because I don’t interact much with a person/band, doesn’t mean it shouldn’t be in my feed.

So. What now?

Good question. One which I’ve been following quite avidly since my previous post about it. People will keep dipping their bread into their soup. Facebook needs to change if it wants to fend off against this impending problem. Either they find a way to beat the 150 number, by changing the way we communicate and stay in touch, or a new social contender comes to the fore that keeps your connections to under 150 (via some implicit social graph machine).

Suffice to say, I am definitely keeping an eye on this! All this talking about soup has made me hungry.

This is a continuation of the API blogging I’m trying to do. I like working with APIs (see Tweekly.fm and TwimeMachine) and I’m currently doing research on social network APIs.

The find of the day is Qwerly.

It’s tagline is: A data API for the social web. It recently ‘pivoted’ a month ago to provide only an API. In short they want to be the “DNS for the social web”. In other words, it is to be the database where social-profiles of people can be found. It started out where one could search for other profiles using a Twitter handle. During April however, the amount of API requests (30 million) far exceeded the 50 000 people visiting the main site. The idea is solid, a service which is lacking. Searching for other profiles isn’t available now, but duckduckgo (an alternative search engine) incorporated it into their search results.

As you can see, it fetched my linked,youtube,posterous, friendfeed and klout accounts. It didn’t find some of my other profiles, because I presume it is because I use different e-mails.

It only pivoted recently and is now almost a year old overall. I foresee definitely great things for them.

P.S. It didn’t find my last.fm account. Considering we (atTweekly.fm) have a db of Twitter+Last.fm connections, maybe they would want to start crawling that as well (need some help)?

About a month ago, I was getting ready to leave for Singapore. I’ve had the privilege of being chosen to participate in SMU’s Apex Global Business IT Case challenge with my university. It was an amazing experience! Now that I’ve caught up with the work I’ve missed, here is blog about my trip.

I’ve only had a great impression of Singapore! The humid weather took some time getting used to, but the rest of the city was amazing. Everything just works, and is clean. It feels like the city is your playground.

The competition itself was awesome. It was so great to meet so many students from so many different countries, all studying what you study.

Our awesome teamhost (Aakriti Goel) took us sightseeing all over Singapore (including the clubs). Without her, our trip would’ve been bland!

All in all, Singapore was an amazing experience, the people I’ve met, the competition itself and the city itself!

Considering it was our first time in South East Asia, we took the next week to visit Kuala Lumpur, Penang and Bangkok. We only had a week, so we opted for going with trains so we can take in the feeling of the countries (not just the cities).

We managed to see the spectacular Petronas Towers and visit the Batu Caves (Hindu religious site within limestone caves).

We then took the train Kuala Lumpur to Penang and then to Bangkok.

Bangkok was crazy. I’ve never seen a city like it before. With our (relatively) laid-back South African cities, Bangkok was a culture shock. By then I was more used to the hot weather, although it still felt like walking through a swimming pool (36 degrees celsius + 90% humidity).

We stayed around the corner from the famous Khaosan road.

When we flew back, we flew via Dubai. I was quite excited to catch a glimpse of the Burj Khalifa. Magnificent testament to human engineering.

If anyone wants to know more about our trip (where we stayed, what trains we took and so forth), don’t hesitate to comment and ask.

As part of my studies this year, I’m doing research into various social network API’s that exist on the web. To motivate me, I’ve thought I’d blog about various interesting things I find: interesting API’s, interesting applications of these API’s and so forth.

Today, I found Google’s Prediction API. This is incredible. It allows you to utilise Google’s machine learning algorithms and get back predictions.

Interesting things that can be done with it:

• Given a user’s past viewing habits, predict what other movies or products a user might like.
• Categorize emails as spam or non-spam.
• Analyze posted comments about your product to determine whether they have a positive or negative tone.
• Guess how much a user might spend on a given day, given his spending history.

It essentially pushes a large chunk of processing to Google. I’ve found very few API’s that actually provide processing. This is new, and could bring forth an interesting trend for specialised processing API’s (that might even charge for its use).

Anyone know of other processing API’s?

It’s been a while since I’ve blogged something. I’m currently doing my honours degree in Socio-Informatics this year and enjoying it a lot! Thanks to my department, I’ve gotten the great opportunity to compete in SMU’s Apex Global Business IT case challenge. In a week I’ll be on the plane to Singapore! I’m going with awesome people and I hope we do well!

I will be taking a lot of photos and will blog the whole trip when I get back.

So. I decided to add permalinks to TwimeMachine along with the ability to view other people’s old tweets.

It stumbled upon some hiccups while developing it, most notably being errors with Twitter’s API and one specific account. It seems that sometimes Twitter says an account has more tweets than there actually are, but I only found it with one account, wikileaks. Now, I’ll be hasty before I jump to conclusions (OMG CONSPIRACY!), Twitter has had errors with tweet counts for quite a while. I tested several accounts (about 20) and only wikileaks seemed to have that error.

Most accounts thus will show the proper amount of old tweets.

So, here is an example of the permalinks:

To read Justin Bieber’s old tweets, simply go to http://www.twimemachine.com/user/justinbieber. To read Lady Gaga’s old tweets, simply go to http://www.twimemachine.com/user/ladygaga. To read it, you must sign up (rate limiting).

As with each new iteration of development, errors might occur and new bugs might pop up. Don’t hesitate to let me know!