Airbnb analysis

Hosts and guests use the same platform but talk about neighborhoods differently. This project mined Airbnb listings and reviews across New York to map the difference.

Role: Data visualization, in a team of five
Product: Interactive website
Timeline: Spring 2017
Links: Project site (external link)

Animated intro of the Airbnb Analysis website — **Fig. 1**The project site, built around four themes: price, interior, identity and neighborhood.

01Context

A neighborhood’s identity is partly marketing. Hosts describe the places they list; guests describe the places they stayed. Both write on the same platform, but do they describe the same neighborhood?

Icons for the four themes: price, interior, identity and neighborhood — **Fig. 2**Four themes to sort the language by: price, interior, identity, neighborhood, plus pace.

02Approach

Text from guest reviews and host listings across the five boroughs was mined, grouped by theme and mapped, so the same word could be compared between hosts and guests, borough by borough.

Diagram: host listings and guest reviews are mined for words, sorted into five themes, then compared borough by borough — **Fig. 3**The method: two voices, five themes, five boroughs.

Side-by-side maps of a word’s frequency in guest reviews and host listings — **Fig. 4**One word, two maps: how often guests and hosts use it, across the city.

Project page explaining the data — **Fig. 5**The data, explained on the site.

Project page explaining the process and code — **Fig. 6**The process and code.

03Outcome

An interactive site that maps the words each group uses most, with word associations by theme. Made with Rachael Dottle, Eric Wong, Clara Dykstra and Andreea Seusan, for Juan Saldarriaga’s data visualization course.

Bars of descriptive words by borough — **Fig. 7**Descriptive words, borough by borough.

Bars of transportation words by borough — **Fig. 8**How people talk about getting around.

Next project

Predicting urban sprawl

Say hello.

Working on something for the built environment? I’d love to hear about it.

LinkedIn Resume (PDF)