This is a periodic newsletter of the interesting things we’ve seen and what we are thinking about in open source policy analysis.

Synthetic data for tax policy featured at Policy Simulation Library (PSL) meeting. Last week, the November PSL meeting hosted by OSPC at AEI featured a presentation from Don Boyd (University at Albany) on a new synthetic data set for tax policy analysis, designed to mimic the characteristics of the pricey IRS Public-Use Microdata Files. Boyd described the steps he and his team took to create and test the data and the data set’s strengths and limitations. Link

New planets discovered with machine learning. One method of discovering exoplanets — planets outside our solar system — is to periodically measure the brightness of the star around which the planet orbits. Every time the orbiting exoplanet passes between the distant star and the earth, scientists can detect a blip in the star’s brightness. One astrophysicist, Anne Dattilo, is using open-source machine learning software to analyze the blips and distinguish between exoplanets and false positives (like “space junk”). In fact, using this technique, Dattilo discovered two new exoplanets this year. Link

PCI-Crackdown introduced at the PSL meeting. At the November PSL meeting, Weifeng Zhong (Mercatus Center) introduced PCI-Crackdown, an algorithm for quantifying the Hong Kong counterprotest intensity. Link

US federal government turns to the open-source community. One advantage of open-source projects is that anyone can pitch in. Numerous federal departments and agencies have moved projects to GitHub and are asking the open-source community for help. Currently on, there are 38 “open tasks” posted by the Consumer Finance Protection Bureau, Department of Defense, Department of Energy, and others. Link

At the intersection of artificial intelligence and econometrics. Most machine learning tools are designed to forecast what will happen next under present conditions. Microsoft’s Automated Learning and Intelligence for Causation and Econometrics (ALICE) software uses machine learning to predict the effects of particular changes in behavior. At its core, the ALICE software helps users measure the effect of an intervention on an outcome while controlling for other factors. The software has a wide range of potential uses, including customer targeting, personalized pricing, and stratification in clinical trials. Link

Chan Zuckerberg Initiative (CZI) announces $5 million in funding for open-source software. As part of CZI’s Essential Open Source Software for Science program, CZI’s first round of funding was awarded to 42 open-source projects, many of which are relied on by OSPC-incubated projects. Applications for the next round of funding open on December 17. Link

Edited by Matt Jensen and Peter Metz