02 Aug 2012

Re: Traffikator

I woke up to this beautiful piece from Bankole: We love Gidi Traffic so much that we're gonna distrupt him. Something about that piece just gave me a sigh of relief. Traffikator is not out, I don't even know the guys behind it, but I love the idea already.

Yeah, I love GidiTraffic too

I don't stay in Lagos. I breeze in every now and then. And even at that, I know how crazy traffic situations can be. Social traffic streams came to solve this. You know them already, you know how they work. And are they solving this problem? I will say they are. Or at least, they are trying. By tweeting crowd-sourced traffic situations, you know what routes to avoid and ones to take. [They tweet more than traffic situations though. I have seen a couple of safety tips among other things].

My (my own) worry

We all have our little daemons. Sometimes, I can really get worried about some little things that are way out of my business. My first question when I noticed one of the traffic stream is who posts this? Really, I was just trying to believe no one will spend a day scouting traffic info and tweeting it. I was wrong.

It's 2012 dear

I mean, why will I want to spend all day scouting traffic info in a technology age that people are planning vacation trips out of earth? What happened to technology and automation? I stand to be corrected, but no one human should post over 350 tweets* in a day. Not by 'listening' to traffic situations in the stream, filtering and 'refurbishing'. No, I don't mean the idea. I mean there should be a better way to automate the process.

It can't be that bad

Automating the process can't be impossible. Difficult? Well, let's say complex. But not impossible. Right here, right now, here are two possible ways I just pulled off head:

  1. One is to write a simple client that watches twitter's public stream within Lagos for certain words like traffic, road and related keywords and auto-retweet. If you want to take a step further, you may want to score each tweet based on the number of matched keywords, position in tweet, word weight and so on and put tweets that fall below a score in a pending list you can then personally vet and approve/reject.
  2. In fact, fuck that. Just find a mapping company with traffic details and implement their API. And yes there are. MapIT, which I met at the last Mobile Web West Africa offer similar solution. The guy in charge even demoed it to me.

Oh yes, traffikator

The guys at Chaos Theory think automation should be it as well. They are putting together Traffikator. And even as just an idea I already love it.

There are humans, there are bots. We should know where we belong.

 

* @gidi_traffic has been up for 11 months (~330 days) and has over 116k tweets. Do the maths.

My name is Opeyemi Obembe. I build things for web and mobile. You should follow me on Twitter (@kehers).