MacroSense ® Technology Platform

MacroSense has been transforming raw location data into actionable intelligence since 2006.

The MacroSense technology platform is central to Sense Networks’ scientific approach to analyzing vast amounts of mobile location data and applying it to useful and engaging products and services.


Mobile location data offers a completely new view of consumer behavior based on places people go.  Sense Networks’ founders – leading computer science professors at MIT and Columbia University – were fascinated by the prospect that one could understand human behavior through looking at location data, understanding the patterns of where people go and how much they move around.

We saw that mobile phone usage was growing at double digit rates every year, and soon everyone would have one. (Remember, 2006 was pre-iPhone!)  Increasingly, these phones give off location data both actively (e.g., Foursquare check-ins) and passively (e.g., app usage).  We saw that this data would provide a way to empirically test hypotheses about human behavior.

Did You Know?

MacroSense processes billions of records in hours compared with days or longer for traditional methods.

Key Features

MacroSense turns massive amounts of mobile location data into actionable, predictive behavioral data. In fact, we currently process 170 billion location points per month into profiles. More than any company other than Google or Facebook.


Powerful Analytics through Machine Learning

Sense Networks has created powerful analytical models that are applied at each step in the data flow. From spatial data cleaning algorithms to identify and correct errors in raw location data streams to machine learning-driven behavioral analytics to measure and predict where people are and where they will go, MacroSense is tuned to transform raw location data into meaningful intelligence.

Real-time Analysis

MacroSense cleans, processes and analyzes data in real-time, and its powerful algorithms continuously learn as new data arrive. MacroSense’s real-time models are based on state-of-the-art machine learning techniques and are powering ad selection decisions that happen in milliseconds. Plus the platform is highly extendable through our deep integrations with open source tools such as R and Weka (Java machine learning package).


Did You Know?

Some of our earlier products include CitySense™, an app for nightlight discovery, and CabSense™, an app for finding cabs in NYC in real-time. See our recent NY Times coverage on CabSense.

Intelligent Location “Features”

One of the toughest parts of most analytics problems is determining which data to use and how to represent those data as variables or features. This problem is magnified when dealing with spatiotemporal data because the data has no inherent meaning on its own.

We have spent years building and refining our MacroSense FeatureEngine™ and feature set. The FeatureEngine™ evaluates each location point to update features that are used for ad decisions and modeling. Included in the feature set are measurements of exposure to commercial stores and demographic areas and calculation of travel behavior, including distance and entropy, to name a few. We have over 2,000 features that we have designed, tested and deployed in commercial engagements.

Massively Parallel Processing

MacroSense is built on top of Hadoop and is massively parallel – we have deployed it in implementations with hundreds of processing cores. Over the years we have built and rebuilt our platform to address the growing size of location data. For example, we have evolved our TreeServer™ geospatial-lookup-software so that today it looks up billions of locations against 15 million points of interest every day.


Sense Networks also discovered ways that this could be done in a privacy-friendly manner: this location data can be used without personal information to bring users a more relevant mobile browsing experience. Moreover, by extracting interesting insights from location data, then throwing the raw data away, you could get rid of sensitive detail that could make people uncomfortable. (see more on privacy here.)