My Top Three Learnings in Omdena’s AI Challenge To Solve the Chicago Gun Violence Problem
Thinking about joining an Omdena AI challenge? My advise will help you to get the most out of a challenge.
My journey with Omdena started when I came across their social media pages on Facebook and LinkedIn. I was wowed because I am very passionate about AI for Good. In the following, I share my top three learnings.
#1 You work with highly-skilled and motivated individuals
I was really excited to find out how much Omdena had contributed to global good by harnessing Artificial Intelligence, bringing skilled, intelligent and highly motivated individuals together to solve a specific problem ranging from environmental problems such as detecting forest fire outbreaks to avoid disastrous deforestation to social problems such as analyzing social media to prevent violent acts.
You also get exposed to several projects that fall in line with what you love doing.
I applied to be a collaborator to join the Chicago Gun Violence Challenge in partnership with Voice4Impact, because I already had experience working with twitter data and its API’s. I took it as an opportunity to improve my skills and also contribute to global good with AI.
The overall methodology was to collect geolocated tweets specifically from Chicago area related to Gang violence, specifically Gun-related violence, and then train a classifier model that can proactively flag down threatening tweets. Voice4Impact helped to get the datasets.
Prior to starting the challenge, I had already gathered extensive experience in the Data Science field, particularly in Natural Language Processing (which is a sub-branch of Data Science), working with and analyzing Twitter textual data as a Research Assistant in the Community Resilience Lab, the Department of Building and Real Estate. Hong Kong Polytechnic University, Hong Kong.
I had also optimized and modified an open-source tool while actively working on some projects in the research lab for mining geolocated old/backdated tweets from Twitter, with minimal programming knowledge required.
#2 You learn skills in the most effective and fun way
I worked remotely on this challenge with a team of over 30 Machine Learning Engineers, Data Wranglers, and Data Scientists from a variety of backgrounds and nationalities skilled in different fields. So it was quite fun while we also learned from each other. Also, Omdena has a policy of not issuing out tasks directly to individual collaborators. As a collaborator, you are expected to be motivated and actively involved in the challenges, and you share your progress with the team in the various channels concerned with what you’re doing. That way you are able to learn and improve on your skills while in a community. We also had weekly calls where each task leader briefs the team about what has been done in the past week, and what would be done in the coming week in order to achieve the goal.
I was the first task manager for the sentiment analysis team, where we used a python library — Valence Aware Dictionary and Sentiment Reasoner (VADER) to first get the sentiment polarities of each individual tweet, and with that, we were able to get a preliminary understanding of the overall sentiments in the datasets. That gave us a prior knowledge moving forward in our ultimate task to build a classification model which was then taken over by Yang Gao who managed the team till the end of the challenge. She shares her wonderful experience here.
#3 You learn faster through collaboration than alone
Being a part of Omdena has been quite rewarding for me in particular, because I've had the opportunity to connect with great minds from diverse backgrounds on a personal level. I have also learned so much from my team members over the past few weeks. There’s this thing about being active on a project, working with teammates and working on your own. You tend to achieve and learn more collaborating with other people, than working solo.
And Omdena provides just the right environment!!
Want to become an Omdena Collaborator and join one of our tough AI for Good challenges, apply here.
