Evaluating the different sources of error in my 2020 elections model
Hello, I'm Phil Thomas. This is my blog where I write about subjects ranging from economics and public health to my most recent programming projects. Feel free to explore the site and comment! —about, blog
I combine forecasts for federal and state elections to find strategic places to focus in 2020.
I use a model of estimated infections to calculate infection probabilities.
I make a few visualizations that I think are missing from the conversation so far.
I use PostGIS to do some spatial queries on a synthetic population of the US.
Visualizing voter turnout and margins by subgroup over the past ten years.
It looks like North Carolina will be pretty important in this election.
Combining state and federal election results to find where voters are especially influential.
Looking at the cross country growth convergence data in light of the great recession.
I build a few interactive visualizations to communicate the concepts behind network models.
A meta-analysis looking at the relationship between intervention uncertainty and impact.
I use the voting power index to find counties that are especially powerful in general elections.
I manually build analytics requests for better page speed and more privacy.
A new project that uses a tabula recta to create and store passwords.
I look at the relationship between risk and expected return in a number of datasets.
I demonstrate a technique for easily using version control along with Jupyter Notebooks.
I show a way to draw numbers from arbitrary joint probability distributions.
A model and d3.js visualization to communicate my thoughts about how the world works.
An interactive, draggable visualization showing turnout, party margins and more.
I compare the turnout for the Women's March and Tea Party by city and state.
I look at potentially close seats in the House and Senate for 2018.
I build an interactive bubble chart to show trends in the 2000-2016 election results by county.
I look at a few graphics showing the election results and build a few of my own.
I take a look at the relationship between social services spending, health spending and DALY burdens across the OECD.
I re-create the polar climate change visualization that went viral using d3.js.
I show a technique for making a stacked chart using the Pandas pivot function
I look at the potential impact of including pandemic disease and a longer life expectancy into the GBD estimates.
Looking at the relationship between individual characteristics and the responses for the economics expert panel
I explore the data from the economics expert panel, focusing on the confidence and consensus among the responses.
An application for checking Gmail via SMS message, built with Python and Twilio
A few tips on some difficult D3.js concepts
Weighing the costs and benefits of blogging
A few ground rules for the blog before getting started