Signing up for my first Hackathon was nerve-wracking. I’d heard that Hackathons are a great way to learn the real struggles of a Data Science workflow, and get some relevant experience, but clicking that “Register” button was a whole other thing! In the Data Science Melbourne Datathon, I would be committing to two months of working with a team of people I hadn’t yet met, building towards a goal that at times would look insurmountable.
At the time, I searched around for advice to prepare me – my Google search history is full of queries like “How to survive a Hackathon”, “What’s it like participating in a Hackathon” and “How to win a Hackathon”. Now that I’m just 3 weeks away from the submission due date, I thought it was about time to give back to the community and share my own experiences in “My First Hackathon”!
What is the Data Science Melbourne Datathon?
From the Data Science Melbourne Datathon website, “What is a Datathon?”:
You work for an analytics consultancy that is pitching to a client for a major piece of work. The client collects data as a by-product of its operations and wants to see if any business value can be extracted from it. You have been given 2 months to demonstrate the potential usefulness of the data and put together findings to present to the client.
This year, the data providers are Public Transport Victoria (PTV) and the data-set is a massive collection of Myki touch-on’s and touch-off’s for trains, trams and buses in Victoria. The possibilities for analysis are endless, and the effects of the analysis could have a huge impact on life in Melbourne, which is really exciting!
The Datathon started on the 24th of July and will run for a total of 2 months. There were two Hackathon days, one on Saturday the 28th of July, and the second was last Sunday (the 26th of September). The Hackathon days ran from 9am until 5pm and included tutorials, talks and pizza – along with a lot of hard work chipping away at our submissions!
Getting a team together
The team at DSM provided us with plenty of opportunities to team up for the Hackathon, including a “team-finder” chat-room and networking time and space at all of the events so far. My team started out of the chat-room.
I joined the chat-room with the hope of joining in with someone else’s team, but I quickly found that most people in the chat-room were in the same boat, and there weren’t many team managers actively replying and recruiting people. So I decided to switch speed and begin recruiting people for my own team. We ended up with Team Girlvsdata – an all women team with a diverse range of career and education backgrounds!
We started off the team with 8 members (pretty much the maximum!), and have ended up with 3 who have committed all the way through.
I’m a Data Analyst, working full time and studying degrees in Business and IT (Data Analysis) at USQ. I program in: Python, VBA and SQL.
A Business Analyst with 5+ years experience. Maya moved to Melbourne from Mumbai about 12 months ago to pursue a career in Data Science. She is currently studying her Masters in Data Science at RMIT. Maya is dependable and hard-working, and an invaluable team member. Programs mainly in: R, Python and SQL.
Has a Bachelors in Biochemistry and a working background in Biomedical Science. She is returning to the workforce after spending the last 10 years raising her 2 wonderful daughters, and we have no doubt she is a great role model to the both of them – she certainly is a great role model to us! Christine’s strengths come from her research background and experience with experimental design. She’s adept at breaking up a problem into smaller chunks and getting into the “why” and “how.” Christine programs in Python and is an incredibly fast learner!
In the first week or so of the competition, I was named team leader. However, team management in Team Girlvsdata has turned out to be a shared task. Our team members help out and volunteer to organize meeting rooms, prepare topics for discussion and share resources like Trello and GitHub (more info on our resources below).
I’ve found that the best way to keep us all motivated is to plan our next meetup, then work towards that. That way everyone has a goal and a timeline. We’ve recently discovered that pair-programming is a great way to get things done quickly. It’s all the benefits of shared knowledge without the inevitable run on time of big group meetings (yay efficiency!).
Tools for Team Management
- GitHub – Great for working on code together and using each other’s code as head-starts on our own mini projects. One of us will generally set up some data cleansing steps in a Jupyter Notebook so the rest of us can skip that step and continue with analysis!
- Trello – Trello is like the digital version of a whiteboard covered in Post-It notes. We are using Trello to come up with ideas, share resources and set up to-do lists.
- Slack – The bulk of our work so far has happened in Slack, discussing ideas, problems and upcoming team meetings. We also use it to share the Meetups that we are attending and other helpful resources we’ve come across.
- Google Calendar – We’ve got a calendar set up for the team so we can synchronize our plans and get the maximum amount of face-time.
What I’ve learnt so far
(The good and the bad)
You will learn a lot very quickly. I’ve already learnt tonnes about working with Python, working with a large team, networking, GitHub, and Data Science workflows.
How to survive a Hackathon day
The DSM Datathon has 2 Hackathon days, which are structured pretty similarly to a traditional short-and-sweet Hackathon. Basically you have about 8 hours to get as much work done on your submission as possible, eat pizza and network.
Just last weekend we finished Hackathon Day 2, which is the last Hackathon day for the DSM Datathon, and I think I’ve just about got them figured out. I’ve got two major points about Hackathon days, keep this in mind if you’re attending your first one (or if you’ve been having trouble with them):
- Hackathon days are paradoxical; you need to get a lot done in a small amount of time, but you’re around a big group of people you just met in a loud room with a tonne of distractions.
- Hackathon days are high-stress events, and not everyone’s going to cope with that in a positive way. You just need to try and keep everyone’s spirits up and get some work done.
How to survive? – Figure out what everyone’s working on early in the day and only ever work with 1-2 (3 max) people at a time. If you really have a hard time getting work done in a loud room, make sure you bring some noise-cancelling headphones and communicate that to your team. There’s nothing wrong with finding a quiet corner and getting some work done, as long as you keep in touch with your team members and make sure you’re on the same page.
Would I recommend a DSM Datathon as a first Hackathon?
I’ve really jumped right into the deep end with choosing the DSM Datathon as my first Hackathon. While most Hackathons are a sprint to the finish line and are over in 1-2 days, (often just 24 hours!) the DSM Datathon is a marathon that lasts for 2 months. Despite the incredible challenge, would I recommend it to other first-timers? Absolutely.
Phil Brierley and the DSM team have gone to a lot of effort to structure this event into one that encourages creativity, learning and independence. Even if you don’t make it to submission day, or you get a submission in and it doesn’t get chosen, you will definitely learn something. My advice is, if you have a passion for data and learning, and want to build a network of like-minded individuals, this is certainly a great place to start!
Stay tuned for Part 2…
Stay tuned for my Part 2 follow up after the competition is over!