A leading geospatial adtech company: Location-based insights from a trillion data points
Companies rely on consumer data to inform their marketing plans, and online activity is only part of the picture. That’s why our client uses location data metrics to deliver information on consumers’ involvement with the physical world. With this complete picture of consumer behavior, their business clients can make intelligent decisions regarding media campaigns and other advertising strategies.
Gathering location data involves tracking hundreds of thousands of digital interactions per second, overlaying this data on top of millions of geographical polygons, and using advanced algorithms to determine behavioral characteristics. Lineate created a data mart and data science testing platform that allows our client to continually refine their algorithms and monitor the effect of these algorithms on their clients’ businesses. We also built end user–facing tools, such as advanced taxonomy mapping, which allow their customers to create relationships between their own data and the information gleaned from the platform.
Share:
Service:
- Data Integration, Analytics, and Activation
Share:
Service:
- Data Integration, Analytics, and Activation
Problem
The company’s team of data scientists needed aggregated data from first- and third-party sources in one place to test their hypotheses, and decision-makers needed a unified point to run reports across a spectrum of data providers to get the full picture of business performance.
They selected Lineate to create a next-generation ad serving platform to aggregate this data in near real time and provide highly targeted campaigns based on buyer intent.
Solution
Our team created a next-generation ad serving platform that aggregates data in near real time and provides highly targeted campaigns based on buyer intent.
Our team also worked to build end user–facing tools, such as advanced taxonomy mapping, which allows the company’s customers to create relationships between their own data and the information gleaned from their platform.
Share:
Features
a custom bidder that processes hundreds of thousands of bid requests per second on the Google Ad Network
a data mart and a data pipeline that preprocesses and pre-aggregates data before feeding it to Tableau, Snowflake, and other business intelligence tools
a data-science testing platform that allows data scientists to build hypotheses on user behavior and test them against geospatial and behavioral data
a fully automated machine learning pipeline and testing framework using a domain specific language
a place intelligence toolkit that helps commercial customers to build taxonomies, rule sets, and their own geographic intelligence capabilities using the underlying platform
Results
The platform seamlessly feeds the algorithms six to nine billion data points per day (several trillion per year) and delivers targeting information in 60 milliseconds or less. Customers saw their revenue potential increase by 7 times, and their marketing teams can now perform their own custom analytics while incorporating deep location insight.
Tech stack
We used Tableau for data visualization and Aerospike as a high-performance NoSQL database. The core application was built in Java, while Snowflake was employed for cloud-based data warehousing. Data processing was handled by Hadoop, Spark, and Impala, and Cucumber was utilized for behavior-driven development and testing.
