This animation of AirBnB host locations from 2011-2014, presented by Ricardo Bion (data scientist manager at AirBnb) at the EARL Boston conference earlier this week, shows the dramatic growth in properties to rent through the service along with the most common routes of travellers. (You can find the R code that created this animation here.)
How did AirBnb achieve such rapid growth? According to Ricardo, it’s by being a “data-informed company”. One of the first eight hires at AirBnb was a data scientist. And as we’ve noted here before, R is widely used at AirBnb along with other data science tools including Python and Jupyter. In fact, most data scientists there are proficient in more than one tool.
Another driver of this growth is a commitment to knowledge sharing, focusing on the tenets of reproducible research, quality, consumability, discovery and learning. Low-quality repositories lead to teams that only read and trust research they have created themselves. To address this, AirBnb created an internal knowledge repostory where data scientists can share “knowledge posts” (research reports created as Jupyter Notebooks, R Markdown notebooks, or Markdown files with all associated code and data for reproducibility). Peer review is accomplished using Github’s issue and pull request workflows, and posts are tagged for easy discoverabilty. In this way, AirBnb seeks to create a central repository of high-quality research trusted by everyone at BnB.
Medium: Scaling Knowledge at Airbnb