DUBLIN DISCOVERIES

TEAM PICTURE!!!

Inspiration

Data is everywhere. Every online transaction we make, every text we send, every map route we search up. Data always plays a crucial part in our everyday lives. As such, we were always fascinated by the idea of having this much information at our disposal, and being able to work to analyze datasets was always something that intrigued us. It was then that we dove into the dataset “StrataScratch: Market Analysis in Dublin,” in order to learn more and help about our Airbnb guests and hosts.

What it does & How we built it

Working in a team of four, we saw ourselves as a diverse group of students with different skill sets and backgrounds. Despite this, we believe that diversity is what fuels growth and success, and our Datathon project proved just that. Our different perspectives not only strengthened our analyses but also ensured comprehensive insight and innovative solutions. Using software such as DeepNote in addition to technologies such as Alteryx Designer to clean datasets and Python libraries (Pandas, Matplotlib, Seaborn, Numpy, Scikit-Learn) to create visualizations, we were able to make meaningful discoveries, recognize patterns, and form accurate observations regarding this dataset. More importantly, we were all able to make contributions to this project and learn something new from interacting with the tools used.

Challenges we ran into

Along the way, perhaps the greatest problem we faced occurred during the brainstorming phase when all we saw initially were numbers and letters in cells looking back at us through the screen. Our main issue dealt with trying to build connections with the information and variables we were provided. As such, we decided to start small with generic bar plots and eventually worked up to more complex visualizations, such as a heatmap and confusion matrix, to better understand the data. At times, we also faced roadblocks which resulted in having to start over from square one. For example, the heatmap took 5 attempts and the confusion matrix took 3. This proved to us that when facing a challenge our team continues to inspire each other in the learning process, concluding that we work and learn best as a team and from each other.

Accomplishments that we're proud of

Throughout our process, we made several accomplishments consisting of finding the geographical search location vs. country, the timing of check-ins given the day and month, the searched neighborhoods vs. price, etc., and using visualizations such as heatmaps, a confusion matrix, world maps, bar graphs, box plots, scatter plots, and distributions. Specifically, we are most proud of the heatmap that describes the Days and Months vs the Check-Ins on those days. Furthermore, we considered many important factors and aspects of our project data to create several suggestions that hosts in Dublin, Ireland can explore to maximize their Airbnb booking.

What we learned

We learned how to work with time series data to create different graphs in Seaborn and Matplotlib. We learned how important it is to continually ask questions about the data and its patterns. Otherwise, we may go down rabbit holes that will lead to insignificant results (e.g. low acceptance rate for guests from India).

What's next???

Our next steps for DUBLIN DISCOVERIES consist of diving deeper into the dataset to discover more connections among the variables. As mentioned previously, we performed trial and error when coding our statistical analysis, since at times, the output was 1) irrelevant to our research and/or 2) not what we expected. As a result, we would have to restart each task, brainstorm, and execute the new relationships that could be formed.