Market Analysis in Dublin

Overview
More Info on Bookins
Hosts be Ghosts
Specific Searches in Dublin
Searches, Inquires, and Bookings
Origin Countries and Interested Price

The Problem

With Ireland’s flourishing landscape characterized by lush countrysides and rugged coastlines, historical sites that offer a glimpse into a deep and storied past that spans from its Celtic origins to its English colonization and struggle for independence, and unique cultural heritage of folklore, music, dance, and literature, tourism is a flourishing sector in the Irish economy. Airbnb plays a vital role in this sector by providing housing for both tourists and citizens of the country. Thus, we hoped to analyze how we could improve the booking experience of guests in Ireland based on what they are searching for in an Airbnb.

Overview

Using Airbnb search data (search parameters guests used) and inquiry data (guest inquiries about listings) from over the span of 11 months in 2014 and 2015, we investigated what guests are looking for in an Airbnb listing. In addition, we investigated how to increase the number of quality hosts to try to improve the overall guest experience on the Airbnb website.

How we built it

We created our project on Deepnote and used SQL, Pandas, NumPy, and Excel for Data Cleaning and Validation. Using these tools, we identified null values in both of our datasets and investigated these nulls to confirm that they had significant meaning and were not indicators of bad data. Further investigation showed strange formatting in filter_room_types in our search data, where the room types users filtered by were all placed together into one column per search query, so we split up these rows and removed null rows to obtain a data frame that could be used to analyze room types users were searching for.

Regarding Data Visualization, such as tables and graphs, we used Tableau, Deepnote Visualization of SQL Queries, and Plotly. In addition, we used Geokeo to retrieve the names of neighborhoods for reverse geocoding our neighborhood codes into neighborhood names, so that we could visualize these on a map.

Using SQL, we found that there were 604 instances where hosts did not reply to guest inquiries, which was indicative of a “bad host”. Diving deeper, we found that of 1184 total unique hosts in Dublin, 273 hosts did not reply to at least 1 inquiry, with a few hosts (5 total) with a relatively large number of instances of ghosting guest inquiries that span to numbers of 10+. We also investigated characteristics of these hosts themselves: the vast majority of hosts only list 1 property at 81.4% while only 5.7% of hosts list 3 or more properties. There are a select few hosts (8) that list more than 5 properties.

Regarding the guests, we noticed that guest bookings peaked in October. A possible explanation is that there are a lot of festivals in Ireland in October, demonstrating that the Irish cultural heritage is a strong driver of tourism in Ireland. We also found the the highest number of people who reserved Airbnbs in Dublin are from Ireland, which demonstrates that the people who want to reserve an Airbnb will most likely already live near the region. We also found that 131 different countries looked at Airbnbs in Dublin out of the 195 countries in the world. This showcases that, as Airbnb city managers, we need to keep in mind the different cultures and traditions that our guests will want to bring with them. Hosts should keep that in mind and possibly incorporate that idea in their Airbnbs.

Interestingly, even though Irish people make the most searches for Airbnb listings in Dublin, they had a lower ratio of bookings and acceptances to inquiries compared to other countries. This may be due to the fact that Irish people already live in Ireland, so simply traveling home the same day may be an option whereas this is not possible for foreigners. Alternatively, hosts may be more inclined to accept inquiries from foreigners because they are willing to pay more for their bookings.

Challenges we ran into

Since this was the team’s first datathon, we definitely ran into some challenges. For some, it was their first time using tools like SQL, Pandas, and Tableau. Although, the biggest challenge we faced was data cleaning. Some members were more experienced with SQL, which made data cleaning easier but it was still difficult because of how big the data was and how we had to learn pandas for the first time to run through the data. We had a lot of firsts during the Datathon and with hard work, we were able to overcome them.

Accomplishments that we're proud of

We are proud of learning how to use many different tools for the first time. Almost everything we used was new to at least one member of the team. It ranged from using pandas for the first time to making density maps in Tableau for the first time.

What we learned

In addition to learning how to use many different tools, we learned how to apply data to real-world problems, specifically in the market. We were able to act like Data Analysts to analyze the best ways to promote Airbnbs in Ireland and around the world.

What's Next for Market Analysis in Dublin

Moving forward, we would hope to discover more associations between groups. For example, associations between the relation between origin of country and bookings. We did analyze many different parts of the data, but in the future, we hope to analyze even more and discover more associations.