Zeenk Releases TQL And Causmos
Zeenk released its Time Query Language (TQL) and Causmos™ causal modeling package. The ecommerce analytics company will make these core pieces of their data manipulation and causal modeling technology freely available to the data science and analytics research community
Zeenk, an ecommerce analytics company, announces that it has released its Time Query Language (TQL) and Causmos causal modeling package as open source for noncommercial research and commercial evaluation. TQL provides a Python based modeling language for transforming time sequence data for analysis and data model. Time series data is central to most business analytics and data science modeling problems and TQL creates a transparent, declarative process for working with the data using the power of Apache Spark™. Pairing TQL with Causmos, Zeenk’s open source library, makes it easy to do different standard causal analyses with just a change in parameters. This software package forms a core part of Zeenk’s current ecommerce analytics solution.
Recommended AI News: IAS Acquires AI Company, Context, to Further Enhance Image and Video Classification
“TQL and Causmos were developed as part of Zeenk’s ecommerce solution to provide marketers with online predictive analytics and incremental value analysis for their Shopify and Amazon programs,” said Brian Eberman CEO of Zeenk, “We are pleased to make this technology available as oen source to the data scientist and research community. This will allow them to accelerate the speed of analysis for a wide range of time-series analysis problems.
Zeenk’s TQL makes it easy to manipulate massive amounts of log, CDP, event database, or other fine-grained, time-series data for data science modeling and analysis. SQL does not provide easy native support for time-based queries. This makes it difficult to develop and scale production analytics and decision systems that support actionable business insights on this class of data
TQL provides a SQL like declarative language making it easy to understand, inspect, and change the assumptions and computational steps used in the analysis. User defined functions are supported for complex modeling tasks. TQL can scale to process billions of records for daily analytics and modeling.
Causmos uses the power of TQL to create causal modeling training and test sets for modeling longitudinal heterogeneous treatment effects (LHTE). Causmos supports models for pre-post, uplift, difference-in-differences, AB-test/randomized controlled trials, and their longitudinal generalizations in both discrete and continuous time (e.g., daily, event-level). Modeling software, such as H20™, TensorFlow™, and PyTorch™, can then be used for estimating causal modeling parameters.
Causmos was authored by Randall Lewis, Ph.D., then a research scientist at Zeenk. Randall joined Zeenk from Netflix to implement his approaches to causal modeling on top of TQL. “Building credible causal models for measurement and prediction is very difficult, Causmos lets a data scientist quickly iterate from end-to-end when building causal models,” said Lewis
Within Zeenk’s ecommerce solution, TQL and Causmos enable marketers to easily access deep insights on their Shopify and Amazon programs with greater speed and accuracy. The solution uses the integrated view from its pixel, third-party advertising, and third-party sales data to provide insights to customers integrated into the overall Zeenk business intelligence experience. Insights can be easily exported and acted on by customers, creating unique revenue and profit opportunities. “The TQL and integrated Apache Spark solution allow us to generate unique insights at scale with high performance and reactivity to user requests,” said Claude Denton, CTO of Zeenk.
Recommended AI News: NetImpact Appoints Sharon Jimenez As CGO
[To share your insights with us, please write to firstname.lastname@example.org]