Data mining challenge
The Data Mining Challenge is a competition specifically designed for data analytics students, and sponsored by SAS, a world leader in analytics software. It is a valuable opportunity to practise the skills we learned, and to learn from the different leaders in the industry. We started by constructing a problem-statement for our target company, which was a movie streaming company. Our team needed to investigate the background thoroughly, not limited to data perspective but also the industry situation and other market information. We aimed to find a flexible and workable data solution using the company database. One of the remarkable solutions was a movie recommendation system. We applied sequential analysis and path analysis to predict customer next movie preference. This could improve the customer experience as the ultimate goal. The data solution was not the end of the path. We also suggested some useful strategies such as reducing the trial period from 30 days to 14 days to improve the cost-benefit control. No pay, no gain. The process of the challenge was not easy, since the data was not ideal for our predictive models. Each team member made extraordinary efforts to create this amazing deliverable.
Enjoying the process
Throughout my university life, I have participated in over five competitions. Most of them, I lost in the first round. It sounds really discouraging. But I learnt from my failure every time. Reflect and improve – that’s the key. Eventually, I won some remarkable competitions such as first runner up of ACCA business competition, and champion of the data mining challenge award. Most importantly, you can enjoy the process through applying the knowledge learnt from lectures and mixing your innovation and team work to create a masterpiece.
(Written by Hanson Chiu. March 2017)