CEU Electronic Theses and Dissertations, 2020
| Author | Balogh, Eszter |
|---|---|
| Title | Identifying Visitors of a Real Estate Portal with a Higher Probability of Posting a Paid Ad |
| Summary | The aim of the project is building a predictive model which is capable of identifying visitors to one of the Hungarian real estate portals with a higher probability of posting a paid advertisement. The data used for the analysis is a Google Analytics clickstream data containing information on more than 100,000 visitors' activity. The predictive model is built with the aim to identify the sellers with a higher probability of switching to a paid plan. A key outcome of the project is that the clickstream data largely available for the company proved to be useful for identifying the target group. The machine learning model built on the data performed well on the test set achieving over 0.91 AUROC value and the variance importances extracted from it are in line with the results of the bivariate analysis. It can be concluded that visitors belonging to the target group are in general more active and also visit the website more frequently. Furthermore, they prefer using desktop and tend to save the interesting ads more often than contacting its seller and they, on average, view properties on sale with a higher price. |
| Supervisor | Békés, Gábor |
| Department | Economics MSc |
| Full text | https://www.etd.ceu.edu/2020/balogh_eszter.pdf |
Visit the CEU Library.
© 2007-2021, Central European University