## How to Re-Map Many Columns in a Database

Introduction A surprisingly tricky problem in doing data science or analytics in the database are situations where one has to re-map a large number of columns. This occurs, for example, in the vtreat data preparation system. In the vtreat case, a large number of the variable encodings reduce to table-lookup […]

## xicor for Confusion Matrices

We have found that for 2 by 2 confusion matrices (a common summary relating the relation between categorical variables) the expected value of the xicor coefficient of correlation specializes into the re-normalized square of the determinant! One can summarize how a 0/1 variable x relates to a 0/1 variable y […]

## Exploring the XI Correlation Coefficient

Nina Zumel Recently, we’ve been reading about a new correlation coefficient, $$\xi$$ (“xi”), which was introduced by Professor Sourav Chatterjee in his paper, “A New Coefficient of Correlation”. The $$\xi$$ coefficient has the following properties: If $$y$$ is a function of $$x$$, then $$\xi$$ goes to 1 asymptotically as $$n$$ […]

## How to Read Sourav Chatterjee’s Basic XICOR Definition

Introduction Professor Sourav Chatterjee recently published a new coefficient of correlation called XICOR (refs: JASA, R package, Arxiv, Hacker News, and a Python package (different author)). The basic formula (in the tie-free case) is: Take X and Y as n-vectors of observations of random variable. Compute the ranks r(i) of […]

## Data Algebra Method Catalog

The data algebra is a system for specifying data transformations in Pandas or SQL databases. To use it, we advise checking out the README and introduction. These document what data operators are the basis of data algebra transformation construction and composition. I have now added a catalog of what expression […]

## Machine Learning Data Preparation in the Database with vtreat

Machine learning “in the database” (including systems such as Spark) is an increasingly popular topic. And where there is machine learning, there is a need for data preparation. Many machine learning algorithms expect all data to be numeric without missing values. vtreat is a package (available for Python or for […]

## Composing Queries in the Data Algebra

When working with multiple data tables we often need to know how for a given set of keys, how many instances of rows each table has. I would like to use such an example in Python as yet another introduction to the data algebra (an alternative to direct Pandas or […]

## What Shape Minimizes Cable Tension?

For no good reason I decided to work out what shape minimized the tension at the attachment points of a draped cable. It turns out to be a lot droopier than one might expect. All of the details of the calculation using sympy can be found here.

## Effect Sizes Say More Than Standard Scores

I’d like to write a bit about measuring effect sizes and Cohen’s d. Introduction For our note let’s settle on a single simple example problem. We have two samples of real numbers a_1, …, a_n and b_1, …, b_n. All the a_i are mutually exchangeable or generated by an independent […]

## Data Algebra 0.9.0 Release

I am pleased to announce the 0.9.0 release of the data algebra. The data algebra is realization of the Codd relational algebra for data in written in terms of Python method chaining. It allows the concise clear specification of useful data transforms. Some examples can be found here. Benefits include […]