457 results for search: https:/www.hab.cl/buy-aciphex-baikal-pharmacycom-rtlx/feed/rss2/copyright
HRDAG Welcomes New Staff, Interns and Fellow
HRDAG is delighted to announce five additions to our team: one new staff member, three summer interns, and one fellow.
Pulling Back the Curtain on LLMs & Policing Data
Structural Zero Issue 04
September 30, 2025
Artificial intelligence is transforming how we work with information. At HRDAG, that changes how I do my job every day. My most recent project was using LLMs to explore and parse vast quantities of data about police abuses in California.
In this newsletter, I’ll pull back the curtain on that work. I’ll describe how a diverse coalition gathered more than a million pages of documents about police misconduct in California and how LLMs helped us make sense of them in ways that wouldn’t have been possible before the advent of this technology.
In addition to understanding my work, I hope that this ...
Reflections: It Began In Bogotá
It was July of 2006, I’d spent five years working at a local human rights NGO in Bogotá, and I had reached retirement age. But then a whole new world opened up for me to discover. Tamy Guberek, then HRDAG Latin America coordinator, whom I had met at the NGO, approached me about becoming part of the HRDAG Colombia team as a research/administrative assistant. Over a cup of suitably Colombian coffee, the deal was quickly "signed.” My responsibilities ranged from fundraising to translations, from support in data gathering for estimates on homicides and disappearances in various regions of Colombia to editorial support to different Benetech-HRDAG ...
Happy Hacking
From my first introduction to the HRDAG community at the annual retreat it was clear to me that mentorship is an organizational priority and that the contributions of interns are valued. Much of my first couple weeks as a summer intern at HRDAG were spent familiarizing myself with Patrick’s paradigm for principled data processing. At the same time, I was learning the skills and tricks (bash, make, vim, git) that promote an effortless programming workflow, a pursuit that Patrick calls “sharpening the saw” (just like in programming, you can cut down a tree with a dull blade, but your life will be much easier if you take the time to sharpen ...
String matching for governorate information in unstructured text
code{white-space: pre;}
pre:not([class]) {
background-color: white;
}
h1 {
font-size: 34px;
}
h1.title {
font-size: 38px;
}
h2 {
font-size: 30px;
}
h3 {
font-size: 24px;
}
h4 {
font-size: 18px;
}
h5 {
font-size: 16px;
}
h6 {
font-size: 12px;
}
.table th:not([align]) {
text-align: left;
}
.main-container {
max-width: 940px;
margin-left: auto;
margin-right: auto;
}
code {
color: inherit;
background-color: rgba(0, 0, 0, 0.04);
}
img {
max-width:100%;
height: auto;
}
.tabbed-pane {
padding-top: 12px;
}
.html-widget {
margin-bottom: 20px;
}
button.code-foldin...
Core Concepts
Inaccurate statistics can damage the credibility of human rights claims—and that's why we strive to ensure that statistics about human rights violations are generated with as much rigor and are as scientifically accurate as possible.
But, what are the pitfalls leading to inaccuracy—when, where, and how do data become compromised? How are patterns biased by having only partial data? And what are the best scientific methods for collecting, managing, processing and analyzing data?
Here are the data pitfalls that HRDAG has identified, as well as some of our approaches for meeting these challenges. We believe that human rights researchers must take ...
Data coding and inter-rater reliability (IRR)
Data coding is the process of converting unstructured information, such as a narrative testimony, into discrete facts such as names and roles of actors (victims, witnesses, perpetrators) in crimes, as well as the date and place of act. Data coding must not discard or distort information. When more than one person is identifying, classifying and counting the elements reported in a qualitative source, the results of what they find may differ slightly based on each individual's interpretation and care in doing the coding. These differences can be measured by measuring IRR (inter-rater reliability). We give the same source document to several coders and ...
New publication in BIOMETRIKA
New paper in Biometrika, co-authored by HRDAG's Kristian Lum and James Johndrow: Theoretical limits of microclustering in record linkage.
About Us
Who We Are
The Human Rights Data Analysis Group is a non-profit, non-partisan organization that applies rigorous science to the analysis of human rights violations around the world. We are a team with expertise in mathematical statistics, computer science, demography, and social science. We are non-partisan—we do not take sides in political or military conflicts, nor do we advocate any particular political party or government policy. However, we are not neutral: we are always in favor of human rights. We support the protections established in the Universal Declaration of Human Rights, the International Covenant on Civil and Political Rights, and ...
Sri Lanka
Ten years after the war ended in Sri Lanka, we still don’t know to the nearest ten thousand how many people perished. The estimates for the death toll for the last five months of the war alone vary between 7000 and 147,000. In 2011, the UN said it thought approximately 40,000 civilians had died; then in 2012 an internal UN report estimated it was at least 70,000. Population data from World Bank and UN sources indicated that more than 100,000 Tamils living in the conflict areas in the north have not returned home after the war.
HRDAG has provided technical assistance to a broad range of non-governmental human rights organizations in Sri ...
Stephen Fienberg 1942-2016
We are saddened by the passing of Steve Fienberg yesterday in Pittsburgh, at the age of 74. He is perhaps best known around the world for bringing statistics to science and public policy and was a beloved professor at Carnegie Mellon University. At HRDAG we are in awe of and grateful for the work Steve did formalizing multiple systems estimation. His work on that front blazed a trail and essentially enabled all of our most important analytical work at the intersection of human rights and statistical science.
If we are to reduce the amount of human violence in the world, the first task is to determine the scope of the violence, to know how much of ...
Estimating the Number of SARS-CoV-2 Infections and the Impact of Mitigation Policies
This Harvard Data Science Review article uses the least unreliable source of pandemic data: reported deaths.
Our Thoughts on the Violence in Charlottesville
This week, we join our friends and colleagues in feeling horrified by the violence in Charlottesville, Virginia. As we have for the past 26 years, we stand with the victims of violence and support human rights and dignity for all. We spend our careers observing and documenting mass political violence across the world. The demands by the so-called “alt-right” to normalize racism and social exclusion are all too familiar to us.
At HRDAG, our work is always guided by the Universal Declaration of Human Rights (UDHR). We reaffirm our commitment to these principles, in particular that the “recognition of the inherent dignity and of the equal and ...
Documents of war: Understanding the Syrian Conflict
Megan Price, Anita Gohdes, and Patrick Ball. 2015. Significance 12, no. 2 (April): 14–19. doi: 10.1111/j.1740-9713.2015.00811.x. © 2015 The Royal Statistical Society. All rights reserved. [online abstract]
Skoll World Forum 2018
Illuminating Data's Dark Side: Big data create conveniences, but we must consider who designs these tools, who benefits from them, and who is left out of the equation.
How Causal Analysis Confirmed Impact of Cash Bail on Verdicts
Causal inference methods show that for indigent clients, money bail increases their likelihood of a guilty conviction.
How Data Analysis Confirmed the Bias in a Family Screening Tool
In Pittsburgh …
BJS Report on Arrest-Related Deaths: True Number Likely Much Greater
(This post is co-authored by Patrick Ball and Kristian Lum.)
Today the Bureau of Justice Statistics (BJS) released a report on their effort to document “all deaths that occur during the process of arrest in the United States.” The analysis estimates that the Arrest-Related Deaths (ARD) program covers only 34-49% of these deaths. A parallel program by the FBI (the Supplementary Homicide Reports, SHR) is estimated to cover approximately the same proportion of deaths. Even taking into consideration both programs, 28% of all police homicides remain unreported.
In order to estimate the total number of homicides that appear on neither the ARD or ...
Primer to Inform Discussions about Bail Reform
The primer addresses what pretrial risk assessment is and what the research supports.