OVER the Easter Holiday early in the year, workers in Kenya’s Nairobi County government ystematically blocked several roundabouts in the capital city. The plan was to prevent motorists from doing right turns, forcing traffic to flow onwards. The eventual plan was to replace the roundabouts with intersections where traffic would then flow in all directions, controlled by traffic lights.
The carnage started the next morning.
As motorists drove onto their usual routes they found barrels filled with concrete, conjoined with metal bars - the usual Nairobi horrible traffic became hellish. It was the first of a host of many measures and the traffic jams and complaints of changed routes escalated over the next few days as Nairobi residents went back to work and school.
“The problem,” says Ma3route startup founder Laban Okune, “is that the county government did not do proper research or simulations.” They simply placed the drums and other forms of barriers and then waited for the chaos to happen - it sums up how little is actually known about Nairobi traffic, an information gap that exists because of lack of sufficient research and investment in proper traffic management and simulation systems.
Laban’s startup, Ma3route, was not initially meant to address this information gap. When he first moved to Nairobi in 2002, traffic was already a mess. When he got to his third year in university, still a neophyte in knowing Nairobi routes, he hit a snag.
He needed to deliver his resume to as many organisations as possible for industrial attachment but more often than not, he got lost or couldn’t find the right matatu - privately owned minibuses used as public transport - to his desired destination. “It hit me then that if there was a resource to pool together information about matatu routes, it would help very many people,” says Laban. It was a simple user problem.
When he got his first car, he joined in the chaos of Nairobi roads. Like everyone else, he blindly followed traffic. Once in a while, if the FM radio updates were timely or some other form of communication alerted him, he based his route on emerging factors. But there was the problem.
Matatu drivers long devised a way to exchange information about traffic. Like any other profession, anything that would lead to opportunity cost was to be avoided, and collaboration was, and still is key. The mode of communication is mainly through hand signals and flashing lights.
The thumbs up means everything is okay with traffic flow, and there are no police officers in the vicinity. Anything else is further defined by specific hand signals. Phone calls, sometimes through the drivers themselves or the conductors also play a key role in this communication.
For private motorists though, this communication channels are completely closed, and they are left to the mercy of the road. When a minor accident blocks traffic flow, the matatus quickly learn about through oncoming traffic and switch routes, leaving private motorists in the mess.
For Laban, this situation was untenable. “What if we had a solution that made it such that as an incident happens, the first people there can help others make decisions?” Such a solution, as Ma3route would eventually become, would follow and perfect the key rules of crowdsourcing. It would automate the process of sharing that information, and find ways to reach as many users as fast as possible.
At the end of 2012, Laban began work on the backend of an Android application. Its primary role, as had dreamt it a little less than a decade earlier, was a resource that provided all the routes followed by matatus in Nairobi. This bore the name of the application and all products that would follow. Ma3route is coined from the sheng’ (Swahili-English Kenyan pidgin) name for matatu, “mathree” which in turn was a simple translation of the last syllables of the word matatu which mean “three.”
As he was building the product, the innovator included a small feature that provided information about traffic on different roads.
One of the key steps towards this massive change was to integrate Twitter with the Android application. Twitter adoption in Kenya escalated in the last 5 years, and its microblogging model made it perfect for crowdsourcing traffic related data. Today their twitter handle, @Ma3route, has over 280,000 followers.
In 2013, Laban presented the app to Pivot East in Uganda - East Africa’s premier mobile startups pitching competition and conference. The startup won the Mobile Utility category. With the spoils came a chance to be incubated at M-lab, and funding that lasted two years. Within that time, the startup grew to a staff of 8 people, and the products increased, and improved.
The incubation period included intensive business sustainability training. The product Laban had developed mostly as a hobby now needed its own revenue streams. He turned to advertising and collaborations with telecommunications companies and application developers. He is also data crunching, using the information collected every day to customise reports for policymakers - for example, seeing how walk bridges along Mombasa Road are distributed. This is seen in how many pedestrians are hit on some stretches of Mombasa Road every year.
The app has also become more intelligent, using user and location based data to give the user customised alerts. Such alerts include the nearest fuel stations and emergency services, especially for users who had just reported accidents. The algorithm for the Twitter account also improved, and can now respond to specific questions such as “How is Thika Road?” The backend is designed in such a way as to include any mentions of stops and other landmarks and bus stops along the specific road.
Another improvement that Laban and his team made during the incubation stages was to find ways to use the data they were receiving. Thousands of tweets are sent to the Ma3Route Twitter account every day, and the account automatically retweets them. The process only sieves out tweets mentioning Alcoblow, a deliberate algorithm customisation due to the moral conundrum of the interventions to curb drunk driving. It also allows for crowd based verification of the accidents, incidents, and normal traffic flow information being shared.
For Laban, one of the most striking elements of the data crunching was how ridiculous peak hours actually are. The applications show a peak between 6-9am, and another between 4-7pm. After midnight, the tweets dip to as low as zero, stretching all the way to 5am when the first tweets emerge.
The data makes sense, but it also provides clear evidence of just how much time and productivity is wasted as people are trying to get to work or home, all at the same time. “The 24 hour economy needs to stop being a dream, “Laban says, “As we are right now, we lose too much productivity just being on the road.”
The spinoff thus became @AccidentsKE, a Twitter handle dedicated to providing analyses on traffic accidents and incidents. In May 2015 alone, 575 accidents were reported through 1,141 reports to Ma3Route. In the first one week of May, there were 136 unique accidents, with most of them involving private cars. Matatus came in at number three, behind trucks. There were fewer accidents involving motorcyclists, and even fewer involving buses and pedestrians.