We wanted to see how our algorithm performed under the interrogation of some of the brightest minds in Data Science.
Working with Edinburgh University’s School of Mathematics we posed two questions:
- What are the key drivers influencing the ‘sold’ price of an item being listed on eBay?
- Can we build a statistical method that takes into account the key drivers, and give the user a realistic and relevant market price for an item?
For the THRIFT team, this project allowed us to engage a diverse group of talent and identify any new concepts or avenues to explore when it comes to developing our system further.
Students used this project as an opportunity to work on masses of real data, for a real problem. Giving them experience in a consultancy setting.
27 of their Masters students selected our 6-week project, we provided them with over 300,000 item listings across various categories and in various conditions. Given the proprietary nature of the THRIFT system, we presented students with data both before and after THRIFTs algorithm cleaned the data.
Over a period of 6 weeks the THRIFT team worked with the lecturers and students to offer support, guidance and answer any questions with regards to the data and possible approaches.
Each student prepared a 3,000 word dissertation which included their conclusion, suggestions and further discussion.
Although the outcome of each dissertation is a secret we will keep, here is some feedback we received:
“Students enjoyed the project with Mark + Vic who were very active, replying to students’ questions, attending meetings, wrangling data etc. At the beginning, it was very difficult and time consuming to set the scope of the project, there were many discussions with Mark and we were able to find a resolution and offer something achievable to students. They have interesting questions and problems and can offer challenging projects. I think this makes it a nice experience for the students.” – Dr Ioannis Papastathopoulos – Academic project supervisor (Lecturer, School of Maths, University of Edinburgh)
“Mark and Vic have been a real pleasure to work with on this student project. They have provided their genuine enthusiasm and valuable time to really work with us to understand how to make the project the best for our students. During the project they were responsive and adaptive and this provided an authentic application of data science to allow our students to work through the problem and demonstrate their skills.“ – Rosie Wilkie (Business Development Executive, School of Maths, University of Edinburgh)
“It was really valuable being able to engage with experts using real world data during the project. Our supervisors were exemplary and really appreciated the high level of energy, engagement and helpfulness” – MSc Stats with Data Science student (School of Maths, University of Edinburgh)
If you’d like to learn more about how THRIFT can be used at your organisation please get in touch.