Mode Analytics Learn SQL Learn Python Gallery Discussions Data Jobs Data News

Facebook's AI factory

Using Machine Learning to Predict Value of Homes On Airbnb
How Airbnb used internal and open-source tools (like Python!) to lower the overall development costs of customer lifetime value (LTV) modeling. Code examples abound. - Airbnb Engineering and Data Science

What’s so hard about histograms?
You’ll never again accept the default bin sizes in a visualization tool once you’ve explored this interactive guide to all the decisions that go into building a histogram. - Aran Lunzer & Amelia McNamara

Inside Facebook’s AI Workshop
When Joaquin Candela first started at Facebook, he worked on an ad-targeting algorithm with a handful of engineers. Five years later, he runs the Applied Machine Learning team, which comprises hundreds of employees running thousands of experiments a day. Here’s how he scaled up Facebook’s AI factory at breakneck speed. - Harvard Business Review

Explaining the Gap: Visualizing One’s Predictions Improves Recall and Comprehension of Data
Something to take into consideration for your next data viz design: this study found that asking people to predict the trend of a line on a chart helps them to better remember the data in the long-run. - UW Interactive Data Lab

Technical Debt in Machine Learning
What do feedback loops, correction cascades, and hobo-features have in common? They’re all machine learning anti-patterns that can slowly creep into your infrastructure and create a ticking time bomb. - Towards Data Science

New from Mode

WrangleConf 2017: Facing bias, ethical obligation, and your audience
Got serious FOMO from looking at the #WrangleConf Twitter feed? Don’t worry, we’ve got a rundown of the best talks on data science ethics from industry leaders like Trey Causey and Sean Taylor.