Technology
Datum's world class technology infrastructure serves as the scalable foundation for its machine learning predictive models including its flagship catastrophic weather predictive application, Eurus®. Our computing capacity and infrastructure rivals those of the largest data intensive organizations, allowing us to mine and pre-process massive amounts of data and train our machine learning models on a continuous basis.
Datum's world class technology infrastructure serves as the scalable foundation for its machine learning predictive models including its flagship catastrophic weather predictive application, Eurus®. Our computing capacity and infrastructure rivals those of the largest data intensive organizations, allowing us to mine and pre-process massive amounts of data and train our machine learning models on a continuous basis.
|
Data Warehouse
The power and flexibility Datum delivers begins with its cutting edge data platform. Our NoSQL Cassandra data warehouse allows for massive and efficient data storage and ease of querying vast data sets. Our data platform currently houses over 10 Terabytes of data. |
|
|
Parallel Computing Power
Cloud based parallel computing enables Datum to produce Big Data insights rapidly and efficiently. We pre-process massive data sets and train our models with Elastic MapReduce across a Hadoop cluster, spinning new servers dynamically as required on the fly. |
|
|
Predictive Modeling
Our state of the art data warehousing coupled with our parallel computing power enable the construction and operation of our predictive models, including our flagship catastrophic weather model, Eurus® Our predictive models use a combination of sophisticated computational statistical strategies to decompress massive data sets into core components to form the basis of our machine learning models to identify complex 3-dimensional data relationships enabling predictive analysis. |
|
