The project

Between March and May 2020, I got an opportunity to engage in volunteer work geared toward solving a real-world problem using Artificial Intelligence (AI). The goal of the task was to build an AI tool able to predict the quantity of food and non-food items (relief package) needed by affected populations during cyclones.

The project was hosted by Omdena and sponsored by World Food Programme. According to its website, Omdena is a collaborative platform to build innovative, ethical, and efficient AI and Data Science solutions to real-world problems. Omdena leverages on the power of collaboration to provide high quality and valuable solutions to the ever-increasing challenges using AI.

When Cyclones hit different parts of the world, they cause devastating effects such as loss of lives, injuries, damage, and loss of properties among other indirect effects. One of the challenges that humanitarian agencies face is to determine the amount of food and non-food items needed during the disaster response. Our project focused on providing a solution to this problem by building an AI tool that would be used by humanitarian actors in response to cyclone disasters.

In writing this article, I aim to share my experience and the lessons I learnt in my role as a Junior Machine Learning Engineer for the project. The team was composed of 34 AI experts and aspiring data scientists. It was a privilege to be part of this amazing team and an honor to accomplish such an impactful project. In this article, I will seek to answer one major question: What lessons did I learn through the project?

Let’s get rolling. Here are some of the lessons I learnt while in this collaborative project;

1. Collaboration has the potential to produce highly creative and innovative AI tools

Data science and machine learning work require a myriad of technical and soft skills. In this project, we had both experts in AI and aspiring data scientists. In addition to that, the team was comprised of people with experience in different areas in machine learning. Ranging from data collection, cleaning, and wrangling to people who are super cool in data visualization and are able to create visually captivating and insightful charts and graphs. In addition, there were those who have mastered the art of drawing intriguing insights from data exploration as well as those who have an edge in feature engineering and creating the models.

These rich and diverse mix of skills and experience developed synergy so well to ensure that the final work was incredible. I learnt that I don’t have to know everything in Data Science to work in a collaborative project. When change makers come together, everyone has something to offer and this is so beautiful. In addition to that, there is a general mental posture of ‘no one is greater’ and the environment is conducive and enabling for collaboration. These complementary skills had a great impact on the final project.

Furthermore, we were people from different parts of the globe and definitely diverse cultural contexts. This too was tremendous. Roy Y. J. Chua, an assistant professor in the Organizational Behavior Unit at Harvard Business School, noted that “To the extent that creativity is about the recombination of existing ideas, then combining ideas that haven’t been connected before creates the potential to produce something new and useful.”

2. Learning is a continuous process and no one ‘knows it all’

No one is self-sufficient as far as AI is concerned. Indeed, no one is an island. I learnt that even the experts in AI were willing and ready to learn and try new things. This was so humbling for me. I was impressed to observe that everyone was willing to listen to other people’s ideas and learn from them. This is because AI is a highly evolving and dynamic field. With the explosion of technology in this age, it is imperative to be ready to learn and grow in the field. This provided a great learning opportunity for aspiring data scientists like myself to easily learn new things.

Other than the technical skills, I was also able to develop my soft skills in problem solving, listening, open-mindedness, empathy and cultural understanding. My worldview was widened even more through this collaboration.

3. Data Quality is a team affair

Now, when handling a machine learning task, sometimes the intrinsic desire is to easily get to the point of writing the machine learning algorithms. This desire can sometimes lead to reduced time spent on ensuring that data is clean and sufficient for the model. So how did we do it?

We worked as a team in making sure that we not only had enough data but also well cleaned. Data collection and preparation was a task for us all. We spent much of our energy together in ensuring that we had enough data in different formats and transforming them to a format that would easily be used for machine learning. The task wasn’t left for a few people to bear the pinch of data cleaning, but all of us worked together on this and the results were satisfying.

4. Communication is the glue that binds collaborative teams

Any project thrives on effective communication. This was one of the lessons I learnt while working on this project. It was good to see everyone communicating the key highlights of the various tasks and the challenges faced. Communication at every step of the project ensures that the whole team is well updated on the project progress. It also helps in tackling complex problems and fostering creativity.

It was interesting to have regular meetings for the team where the summary of work done was shared by the different task groups. It helped keep focus and prioritize on different stages of the project. Moreover, sharing the highlights and celebrating the milestones is a motivating factor in project progress. Throughout the project, effective communication helped nurture openness and develop mutual trust and respect that enabled us work together on different tasks. It also ensured that our collaboration was not just about finishing the tasks but also building a community of change makers.

5. Take up challenges and work hard

During the entire project, I realized that the best way to succeed especially as a learner is to take up challenges and work hard towards providing a solution. In a fast-paced machine learning ecosystem, where there are a lot of tasks to be done, it is good to stretch yourself in taking up even the tasks that look difficult. Do not wait and watch as other people work. Be willing and ready to pick up tasks and work on them. In doing so, you are able to try different ways of solving the problem, read, and research about the tasks. This is of great benefit since at the end of the day, even if you are not able to do it perfectly well, you have learnt the process and the intrigues around it. The mentors were also helpful in helping out with challenging tasks and sharing their work for one to learn.

Be ready to go to the deep ends of machine learning. Don’t be afraid for trying some of the things you have learnt. I enjoyed the freedom and the good will that was accorded everyone to try out ways and techniques in solving a problem. This is a great resource and opportunity for everyone who is yearning to learn and grow in this field. You are free to think outside the box, experiment and deliver results.

6. Document your project journey.

I took it upon myself to document the things I was involved in doing. But not only that, I also kept a record of the new things I needed to learn to be better in my work. In every meeting, I had a pen and a notebook where I scribbled notes about the tasks that we were to do.

There is a way in which writing helps in staying focused on the projects and helps the brain to be alert and on course. I enjoyed this experience. In some cases, I would hear highly technical discussions and I would put them down for future reference and learning.  I saved websites’ URLs that I would later search to find out more about the ideas. This kept me going and was indeed beneficial. When I look back, I appreciate the effort I put into making sure nothing escaped my attention. I was in the learning mode throughout.

Conclusion

Generally, I learnt a lot from this project. I am so grateful to Omdena for giving me the opportunity to work on such an amazing project. I also appreciate the entire team who worked tirelessly to ensure we delivered a quality product. Indeed, you are change makers! Finally, thank you World Food Program Innovation Accelerator for sponsoring this project. We hope that what we did will achieve much more in the humanitarian sector. For more information about the project, read this article, or watch the demo day recording video.