At Opensignal I had access to incredible amounts of data on real world mobile experience. I wrote and published an article that examines the role that various signal strength and temporal factors influence experienced speed. This is an interesting problem because, unlike 3G connections, LTE connections have three different signal strength metrics, and it can be tricky to know which one to focus on.

This article has quietly accumulated a decent number of citations over the years since. My guess would be that academia has limited access to datasets like this one, and so any large-scale empirical studies can be hugely valuable.

This was a great piece of marketing for the company, and we ended up using it in a lot of sales conversations when we wanted to demonstrate our academic credentials. While doing data science in industry is always a balance between scientific rigour and commercial concerns, this was a project where we got the balance right.

In retrospect, I wish I had done more like this, there was enormous potential in both serious scientific research and lighter content marketing with this dataset.

You can find the full article here.