Last week of July more than 6,500 statisticians from 52 countries gathered for the biggest statistical conference in the USA – Join Statistical Meeting. They were discussing new research and ideas during the sessions and talking about values and distributions during lunch. Business was there too, and they were on the hunt for talented statisticians and data scientists.
Tech giants have access to the richest data and are using it to make corporate decisions, but it doesn’t stop there. They are looking to hire statisticians, experts in machine learning and artificial intelligence with diverse backgrounds to put their data in use to tackle research challenges. A lot of them collaborate with academia and government.
Data scientists at Facebook conduct large scale research to understand better how people interact with each other and react to the events in the outside world. One of the examples is analysis of the Facebook disaster data. When a disaster happens, Facebook uses analytics tools to see where people are moving and then collaborates with government agencies to help them with information on where resources are needed at the moment.
Another Facebook working research project with social influence is to analyze what Facebook groups are “healthy” and what are not so much by looking into most active members – the influencers – and into what type of message they send and how other people in the group react to that. The goal is to determine if the group is healthy and if not – what to do with that.
And of course, there is a collaboration with political scientists to look into election activities. The tasks are different for different countries. To figure out what is the best use of resources, they have to talk to local policy makers to see what impact the results of the research can have, what low hanging fruit they can tackle. Because if it doesn’t have a chance to have policy impact, they won’t do it.
Facebook also has a website dedicated to their research results and projects that is available for public access.
LinkedIn is looking for bright minds too. In a similar spirit of collaboration, a lot of research is done to have an impact outside of the company. One of the newer projects is creating economic maps to show where professionals are located. Also, with obviously a lot of information about the labor markets, a big emphasis is on the research of what skills are needed for future employment and what could be taught in schools.
Netflix is well known for using machine learning for creating the sequence of the movies you see on your feed and to improve other user experiences. But there is more – pretty much every single decision in the company is based on the research. Every movie release, the size of the “play” box on the screen, star or thumbs up rating is chosen based on control trials. And when control trials are not possible, Netflix uses quasi-controlled experiments and causal interference.
Similar to Facebook, the overview of the research is published on their own research website that they launched last year.
Not only tech giants are looking to hire, but young companies too. For example, Didi – data driven Chinese transportation company with a mission to become a global leader in transportation. They start with collecting as much data they can, including traffic conditions, driver’s and user’s behavior, and then using big data and machine learning and AI algorithms to improve traffic capacity. One of the very many ways to do it is to use data driven allocation strategies, including traffic and supply and demand forecasting.
And no statistical gathering could be complete without many companies that rely on the golden standard of statistical causal analysis – the clinical trials. Pharmaceutical and medical industry have been in demand for statisticians for decades, and are using everything that’s available in the field – from traditional clinical trials for new products to AI models using machine and deep learning techniques to develop better customer experiences.
From tech companies using big data in decision making to pharmaceutical companies doing clinical trials to Census and National Security agencies making policy decisions, statistics is in demand everywhere. With the amount of data available today, this was a hot place to be.
By Daria Milakhina, PhD, Audit Methodologist, Kansas Legislative Division of Post Audit