Timeline Query Language (TQL)

Zeenk’s Time Query Language (TQL) and Causmos causal modeling package is now available as open source for the data science and analytics research community. Interactively build complex datasets from time series data, using our TQL data-wrangling library.

Enable Measurement of the True Causal Effect of Decisions

Want to know if an ad campaign or pricing change actually caused revenue to go up? To determine the causal effect of your decisions, you need to model the alternative choices. 

Zeenkā€™s Time Query Language (TQL) and open source Causmos library radically reduce modeling time. So while others are struggling with the expensive process of developing even a single causal model, Zeenk enables your teams to carry out multiple studies to support your decisions at a reasonable level of effort.

Traditional Causal Modeling Is Hard

Traditionally, causal modeling can take weeks or months of data engineering. Extrapolating multiple potential outcomes from a single observational data flow. Extracting observational data from log files. Processing and converting it into specific data tables for training and testing. Plus, each different causal technique requires different processing and table information, which takes even more time and effort.

Causal Modeling With TQL Is Easy

With TQL and Causmos, you can simplify modeling and accelerate your teams without the challenges that plague traditional causal modeling. 

Want to rapidly explore many models? TQL can read directly from your raw data and process it into specialized timeline structures. Then your data scientists can use tools they already know, like Jupyter, to declare how to transform these timelines into the training and test files they need. So why waste valuable time on building 1:1 datasets, training models, and comparing results, when you can use TQL to leverage timelines, configure training, and iterate rapidly and at scale?

Pairing TQL With Causmos

Causmos, our open source library, makes it easy to do different standard analyses with just a change in parameters. Together with TQL, your data scientists can use Causmos to quickly test the plausibility of different assumptions, including counterfactuals to inform your business decisions.

Out of the box, Causmos supports a number of standard causal modeling techniques, including:

  • A/B Test
  • Pre-Post
  • Uplift
  • Diff-in-Diff

Because Causmos is open source, you can also use your favorite data modeling techniques from H20, Tensorflow, PyTorch, or other tools directly on data normalized by Causmos and TQL.

Ready to measure the true effect of your business decisions? Contact us to use TQL for free.