SMART (Spatial Monitoring and Reporting Tool) is one of the leading poaching reporting tools in use in protected areas around the world. Effective as a specialist software, it fails to capitalise on input from the communities living in and around conservation areas.

Capturing local community intelligence in a meaningful way could be a game-changer for wildlife, bolstering ecological monitoring and law enforcement efforts. The SMART Partnership are seeking to use technology to this goal. In order to be successful, any such solution should be simple to deploy, easy to use, and able to overcome language, technology and literacy challenges affecting these communities.

Our inspiration stemmed from joint appreciation of elephants and familiarity with the research and awareness work of Save the Elephants in Kenya, who brought Samburu to Google Maps and the attention of the world. We realised we could do so much more with technology. We want to create a harmonious loop:

  1. Report (incidents such as poaching, with the help of engaged and incentivised communities)
  2. Analyse (large datasets to identify circumstantial patterns)
  3. Predict (the likelihood, time and location of criminal activity to help facilitate proactive over reactive law enforcement)

What it does

Rafiki is a concept aimed at engaging communities in the conservation of local protected areas. Rafiki makes it easy to report incidents of criminal activity (such as poaching and illegal trade) and tries to overcome language, literacy and technological challenges. Eventually, Rafiki begins to predict patterns in criminal activity and sends out pre-emptive alerts to rangers and communities.

At the very least, Rafiki is a progressively enhanced solution that works across a variety of channels: SMS, common to feature and smartphones alike; voice; and smartphone app – localised based on access number, area code and GPS data.

In the case of SMS, the interface is conversational rather than being a prescribed set of questions; natural language processing extracts critical data from text.

In the case of the app, a thoughtful user experience and rich visual language removes the barriers to engagement and educates communities with video and audio snippets, providing a compelling call to action.

All in all, Rafiki makes it easy for those living in or around protected areas to report incidents of critical concern to conservation, safely (and optionally, anonymously) and just in time.

But it doesn't stop here; we identified the potential for using external datasets (weather, seasons and daylight, animal migration patterns, camera traps, historical incident reports, news headlines) & machine learning to offer predictions of when, where and how criminal activity is likely to occur. Early notifications could mobilise law enforcement, instil community vigilance (and prompt re-engagement) and may act as a deterrent for criminals.

How we built it

We started with the SMS interface, which is implemented and currently operational. We also prototyped a web app/native application, all based on a microservices architecture.

We built a proof of concept to establish if the machine learning component could be an integral part of the system. The data feeds available are the limitation to the potential of a system such as this. For example, if live tracking data were available (with the obvious protections in place) our machine learning back-end would have been significantly enhanced.

Challenges we ran into

Machine learning is an exceptionally complex area and as such we chose to focus on the key points of the brief. The paucity of available datasets made implementation intractable within the time constraints of the hackathon.

We’re a team of two, so we had to restrict the scope of implementation; however, we did not restrict the project scope and resolved to be prepared to deliver everything we’d propose, eventually ☺

Accomplishments that we're proud of

Rafiki’s cute logo! The harmonious loop: report, analyse, predict. 48 sleepless hours and so much passion. Also it's great to have a fully functional, multilingual application (give the SMS a try!).

What we learned

The Afrikaans word for ‘poaching’ is ‘stropery’, according to Google Translate. Conservation presents many challenges and exciting opportunities for the world of tech.

What's next for Rafiki

Our main focus to come is the machine learning platform, which we believe have enormous potential for law enforcement and community engagement. We extremely excited about the potential this offers and completing further research and implementation around this would be the obvious next step.


Text Rafiki to +44 7400 047859 (English) or +44 7400 274155 (Afrikaans)

Rafiki Presentation

Share this project: