This website uses cookies and similar technologies to understand visitors' experiences. By continuing to use this website, you accept our use of cookies and similar technologies,Terms of Use, and Privacy Policy.

Sep 24 2019 - 02:46 PM
A Data Science Coder's Night-out - Columbia Data Science Hackathon

A popular running joke about software engineers (specifically coders) in India is about how there are times in their lives where he/she will have to work the entire night ("night-out") to write code. One of my few night-out was last Saturday night (Sunday morning).

Yi and I, along with a couple of our friends participated in the Columbia Data Science Society's annual data science hackathon which was held from Saturday (21st) evening to Sunday (22nd) afternoon

The event started at around 7:00PM on Saturday where we were welcomed to the hackathon and a couple of judges (Data Scientists working at Facebook and NYC Data Science academy) spoke about the work their companies were doing in the field of data science. We were then introduced to the dataset for the hackthon. The data given to us was about the daily traffic of customer data for 155 stores owned by 3 grocery store chains (Trader Joe's, Whole Foods Market and Sprouts Farmers Market) in Texas, Colarado and Georgia for Q2 2019. The data also contained various attributes like location of the store, average distance to the store from the home and workpalce of the customers etc. We were also given data on what region the customers going to the stores belonged to

Equiped with this data, we started working on the data from 8 pm onwards and worked through the night. We were able to build Dashboards (on Tableau and Carto), a Bayesian Hierarchical Spatial Model, and a Presention (check out if you are interested). Finally, after 14 hours of work, (with a short dinner break in between) we were able the send the presentation to the judges at 10 AM on Sunday.

At about 11:30 AM, we were announced as one of the 7 finalists who would be getting a chance to present our work to the judges. Though we were tired and sleepy, we managed to present our work to the judges.

Unfortunately, we were not among the top 3 placed winners. But as a consolation, one of the judges personally came to us and expressed his appreciation for our work. Though the judges liked our work, we could not present the work that we did effectively and probably lost out on our presentation. Nevertheless, it was a really great experience to work with the team and we learnt a lot from each other

Our teammates were Akanksha Rajput, a second year master's student at Data Science Institute and Ming Zhong, a second year master's student at Statistics Department. Incidently, Ming was one of the orignal contributers to the "Dog Data Scientist - Engaging Children in Data Science" that was presented by Ipek Ensari and me at the "Data Science in Classroom" held at the learning theater in August..

Team Picture taken after the event (Yes, everybody looks tired and sleepy 😀)

Key Takeaways and learning

  • Brainstroming ideas with teamates can really help develop really good solutions to problems that look tough.
  • "Jo Dikhta hai, wahi bikta hai" in hindi, which loosely translates to "that which is seen is sold". Try to present all the things that you do effectively. Ensure that people are able to see the good work that you do everyday.

Though this night-out coding effected my sleep cycle (Can I call it jetlag?), it was a great experience working with Yi, Akanksha and Ming. Notwithstanding the fact that we returned empty handed from the competition, I can say that this experience would definitely help me in such competitions I attend in the future.

Contact Yi or me if you would like to know more about our work

P. S. Coincidently, my last night-out coding was when I was applying for the position of Summer Intern at Edlab. Time does fly!

Posted in: FYIEvents|By: Ameya Karnad|1719 Reads