My experiments in the world of hackathon started out of my boredom in office work and have turned out to be a rich collection of experiences. You can find the solutions at my github.
RBL bank hackathon(3 days)
Problem statement — Make a data science solution with customer data
Offline | Prize pool — 2 lakhs
What I liked
- Great food arrangement
- Comfy workplace
Could have been better
- Forced to use API provided for fetching data
- The API didn’t work for 1.5 days
- API have transaction data of only 1 user of just a few months. No machine learning was possible over it.
- No guidance provided on what to do with so less data
- Team presentations were private with the Jury
Coinberg hackathon
Coinberg hackathon(2 days)
Problem statement — Cryptocurrency — Arbitrage trading | Sentiment analysis | Portfolio management | Trend forecasting
Offline | Prize pool — 2 lakhs
What I liked
- API given for fetching cryptocurrency data
- Delicious snacks and drinks
- Energetic organisers
- A nice kit containing t-shirt, sketchbook, pen and a fancy pen-drive in a beautiful bag
- Energetic diverse crowd
- Decent sleeping arrangement
- Team presentations were open
Could have been better
- WiFi didn’t work for the first 2 hours due to unexpected load and number of connections. Everyone worked on the mobile internet.
- Result announcement took a week
- I won the competition and the payment came 3 weeks post announcement
Morningstar hackathon
Morningstar hackathon(4 days)
Problem statement — Extract relevant tables and datapoints from annual reports of Indian companies
Offline | Prize pool — 1.5 lakhs
What I liked
- Good dataset
- Personal assistance
Could have been better
- Team presentations were private with the Jury
- Less prize money ( I won the event and the money seemed less for the efforts)
Axis bank hackathon
Axis bank hackathon(2 days)
Problem statement — Extract and normalise tables in given PDF/images
Offline | Prize pool — 5 lakhs
What I liked
- None — Really the worst hackathon ever
Could have been better
- Test data was very different than train data which could have led to the erratic performance of teams. There were 2 problem statements and both had this issue. It disappointed a lot of teams.
- During the first phase, organisers suddenly introduced making a working API. Then later they removed it.
- The organising team seemed unenthusiastic
- Bad food arrangements
- Everyone’s code was tested by the organising team and participants were filtered for next round. This step happened without interaction with participants which can create issues while running code.
- Travel reimbursement took 50 days after several emails from all participants
Microsoft hackathon
Microsoft hackathon(35 days)
Problem statement — QA ranking system — Find the best answer to a question given 10 answers
Online | Prize pool — 10 lakhs to the professional team and 5 lakh to the student team
What I liked
- Ample data
- Microsoft’s Codalab website seems good for organising hackathons
Could have been better
- Too much data required heavy compute for deep learning solution which must have filtered out a lot of people
- Announcement of the result was delayed twice
- This was a difficult problem but the prize was given to only 1st professional and 1st student team. It could have given to first 3 teams in both categories.
Hackathon solutions
https://github.com/bhavsarpratik/hackathons
Originally published at ml-dl.com on February 23, 2019.
