What is Empirical?
Empirical is an analytics engine that automatically models structured, tabular data (such as spreadsheets, tables, or csv files) and allows those models to be queried to uncover statistical insights in data. Our engine is designed for you to interactively explore, explain, predict, and simulate data. Empirical can be called from Python and R, used via web front end, integrated with spreadsheets and BI tools, or embedded into custom applications.
In all of these cases, modeling and querying happen in real-time, so users can interactively work with their data to:
Discover relationships and be guided to insight
Evaluate data quality
Identify anomalies and areas for investigation
Make explainable predictions and test different assumptions
Run what-if scenarios and simulate new data
Empirical is an embedded analytics engine used by software development teams, product managers, and IT executives to power their own human-facing applications with automated analytics and guided insight. The Empirical analytics engine has many capabilities, and when it’s embedded, you decide which ones get exposed to your users, via your own UI.
Empirical as a stand-alone product is used by quantitatively minded people such as data scientists, statisticians, and advanced business analysts who are familiar with statistical concepts like correlation and confounding variables. Empirical comes with its own web UI for data exploration and various APIs.
If you’re a business user, think of Empirical as an automated statistician, delivering quantitative capabilities that may otherwise be out of reach for you. If you’re a data scientist or trained statistician, think of Empirical as automating many of the tedious tasks that you should routinely do.
Empirical’s unique value proposition is augmenting human domain experts with analytic insight into their data. Any application or use case that includes humans looking at structured data and making decisions is a good use case for Empirical. The system works with tabular structured data which is common in databases, spreadsheets, and Business Intelligence tools. Empirical’s modeling approach handles missing data, complex data, and very wide data, which are common in business applications and operations. Common use cases include:
Sales. Understanding factors driving opportunities, benchmarking reps, churn.
Marketing. Dynamic targeting, amplifying controls, optimizing campaigns
Human Resources. Predicting employee retention, expected compensation, recruiting
Product Management. Finding patterns in user behavior, hypothesizing new features
Clinical Trials. Identifying adverse effects, selecting site locations, recruiting doctors
IoT. Understanding device lifecycles, correlating signals, scoring data
Building Management. Identifying similar buildings, running what-if scenarios for capital improvements, simulating occupancy
Empirical is often compared to Auto ML tools and trendy algorithms such as deep learning, where the objective is to automate a simple decision for operational efficiency or cost. These systems usually have big data and a single or limited set of questions to be answered: “What ad might this user click on?” “Is this image a cat?” “What is the likelihood that this transaction is fraudulent?”
However, the majority of business use cases involve small- to medium-sized structured data. Empirical is designed for exactly this kind of data. If the objective is to understand business data so that humans can make better, thoughtful data-driven decisions, then what you need is Empirical.