epogrebnyak / economist-tools

Dataset access, learning resources and research reproducibility

Dataset access

You can migrate your data analysis from Excel to R/pandas or use R/pandas to prepare sunsets of data and work back on it in Excel.

I developed several tools that allow access to useful country and corporate datasets.

• weo - IMF World Economic Outlook python client (pip install weo)
• boo - Rosstat annual corporate reports (pip install boo)
• comtrade - utilities to query UN Comtrade database for goods export

weo and comtrade wrap already good datasets, and comtrade API is nearly perfect. boo makes use of raw data, which is otherwise unaccessible.

On the radar

Here are some code projects by and for economists (to be annotatted):

• https://www.hse.ru/edu/courses/302789733
• https://github.com/WLM1ke/poptimizer
• https://github.com/lnsongxf?tab=stars

Please send in your own similar projects for listing.

Reference and learning

I maintain a website with open-source textbooks and guides in
econometrics and statistics, Econometrics Navigator.

Programming and IT skills

• learn/python - top 3 resources I recommend as quick intro to python
• haskell-intro - a starter guide into functional programming with Haskell
• superhero - business analyst-to-developper learning curriculum (in Russian)

Replication and reproducibility

Economic analysis is not terribly advanced in replication and reproducibility. Required skills taught in fragments and there are powerful incentives to keep your work to yourself, but hope things are changing (open access publishing, literate programming, data APIs, etc.).

I contributed some code to the following open source tools:

• manubot - changes to citation management submodule
• handout - minor fixes

Mantras and opinion

• Analysis is a DAG - try eliminate manual work parts or undocumented operations from your our data transformations.
• No silver bullet - whatever you try does not solve 100% of your problems.

My other pages

• Blog
• CV