ᴰ³: Dublin's Demand-Supply Decoder

Host Response Times Based on Booked & Not Booked Airbnbs
Tapped Neighborhoods Stacked-Bubble Chart
Confusion Matrix Logistic Regression
Conufsion Matrix Gradient Boosting

Inspiration

Our team was inspired to improve the quality of Airbnbs in not only Dublin, Ireland but also create a way to analyze and improve Airbnbs in different host cities using data regarding guest searches and inquiries.

What it does

Upon looking at the data, our team wanted to look deeper into Airbnbs and what market trends lead to more potential customers. We also wanted to see if the connections we made and the analysis we did for the Airbnbs in Dublin could also be applied to other host cities to improve guest demands there as well.

How we built it

We built our project by splitting the data into two different categories and subcategories; guests who made searches but DID NOT make an inquiry to book and guests who made searches and DID make an inquiry to book. We further broke the latter down into guests who made an inquiry and DID book and guests who made an inquiry but DID NOT book. Through analysis of the first category, we were able to find the discrepancies between what guests who made inquiries found and what guests who did not make inquiries failed to find. Through this, we were able to extract information regarding what guests had searched for before choosing not to make an inquiry, showing what categories of the Airbnb selections left guests disappointed the most. By narrowing these gaps, guests would be able to find Airbnbs in Dublin that fit their criteria and hosts would have an increase in inquiries. Through the analysis of the second category, we were able to look deeper into whether host quality and poor communication played a part in guests deciding not to book the Airbnb. With this analysis, we can provide hosts in Dublin with data regarding the aspects of the Airbnb that guests found unfulfilling to help retain bookings after the initial inquiry is made.

Challenges we ran into

One of the challenges we faced when tackling this problem was finding a viable option to combine both data sets and compare them as there was a lack of matching primary and foreign keys connecting the sets. We had planned to create whole new tables with primary and foreign keys that we would later use to analyze the data, but after looking back at the bigger picture, we realized another way in which we could analyze the data tables given to us as is without having to painstakingly create new tables ourselves.

Accomplishments that we're proud of

One of the accomplishments that our team was proud of was that for most of us, it was our first datathon, so even being able to churn out a completed analysis and presentation was an achievement for us. Additionally, we learned a lot during the process of cleaning, analyzing, creating graphs, and validating the data, especially since it was new territory. One of the graphs that we felt accomplished after making was the bubble chart that reflected the most popular neighborhoods in Dublin that were searched for by guests. We ran into a lot of issues when making this graph, but we wanted to push through and create this since it was visually appealing and unique to our presentation. Also, being able to cross-validate after we were done exploring the data was very fulfilling and allowed us to make concluding statements about what we found.

What we learned

We learned to look at the data given to us in a certain way that aided in creating an algorithm that produced shortcomings of Airbnbs in Dublin to improve the number of inquiries and flaws in host quality and communication that could help build and maintain a strong customer base.

What's next for D^3: Dublin's Demand-Supply Decoder

Our next steps in our project collect data from Airbnb from other cities and compare their trends with the ones we made in this project. The comparison would help Airbnb approach each area with its local context, curtailing Airbnb’s algorithms to offer better options for the searches given by the guest. We want to make sure that Airbnb’s guests know all the options that they have and make sure they are getting what they are looking for.