A team of two Civil Engineering Department students under the name "Data Dudes," Will Sanders (ME Student) and Trevor Stull (BS student), won first place in the first university-wide annual Analytics Competition hosted by the Association of Information Technology Professionals (AITP) held in the Department of Information Systems. “Data Dudes” were given 24 hours and over 500,000 rows of data with information on accommodation's latitude and longitude, list pricing for accommodations, and other related data provided by AirBnB.
Trevor Stull employed a spatial analysis that analyzed key attributes of the listings, plotted in over 100 ArcGIS maps using latitude and longitude. The team concluded significant spatial auto-correlation for both the number of reviews in a listing and the price of a listing. Trevor also calculated the coverage area of AirBnB listings over time, which can be used to forecast the spatial growth (ground area) of AirBnB's service.
Will Sanders used a regression analysis that predicted the number of reviews of listing for leveraging the machine learning technique of “random forests.” After hours of testing, tuning, and iteration, Will arrived at model that predicts the number of reviews with Mean Square Error of 95.5 and R2 = 0.917. This prediction is valuable because the number of reviews can be used to estimate number of visits, and number of visits can be used to estimate revenue.
Both Trevor and Will have worked as research assistants for Dr.s Mattingly (PI) and Fang (co-PI) on the North Texas Regional Housing Assessment Project funded by the City of Dallas.