How to get set up with ollama on the Minnesota MSI HPC cluster; from installation to running a model to pulling weights and setting up a model from hugging face.
For a recent revision, I have had to dig into TAQ data; I did not want to use SAS and I thought a combination of command line unix tools and julia would be easy. I learned a few things!
I have been asked how to set up a simple Julia IDE (not VSCode) for people who are already using Sublime Test. There are a few things that need to line up for it to work smoothly. I wrote these steps down.
A guide for moving between Julia’s DataFrames.jl and R data.table. This follows the template of stata2r and gives side by side matches of common operation for data wrangling.
There are (were?) some issues with F-statistics estimation in stata for IV regressions. We simulate data to reproduce the problem and present a fix in the julia FixedEffectModels.jl library.
Some pictures in TikZ for my JF paper on Asset Pricing with Entry and Imperfect Competition: (i) a stylized Hopenhayn model; (ii) a Hotelling-Salop circle; (iii) some demand elasticity curves.
Some pictures in TikZ for our JFE paper on Bubbles and the Value of Innovation: (i) a picture of detailed disagreement (favorite); (ii) standard blocks of an equilibrium model with heterogeneous firms.
Some pictures in TikZ for teaching a standard course on investment: (i) Capital Allocation Line (CAL); (ii) Efficient Frontier and Optimal CAL; (iii) Security Market Line; (iv) the link between the optimal CAL and the SML; (v) Treynor-Black portfolio selection.