rwa | Machine Learning on Sequential Data Using | Machine Learning library

I have a dataset with about 50 columns (all indicators I got from World Bank), Country Code and Year. These 50 columns are not all complete, and I would like to fill in the missing values based on an lm fit for the column for that specific country. For example:

Doing this for a single country and a single column is absolutely fine when following these steps here: Filling NA using linear regression in R

However, I have over 180 different countries I want to do this to. And I want this to work for each indicator per country (so 50 columns total) So in a way, each country and each column would have its own linear regression model that fills out the missing values.

Here is how it looked after I did the steps above: This is the expected output for ONE column. I would like to do this for EVERY column by individual country groups.

However, the data looks like this:

There are numerous countries and columns that I want to perform this on just like the post above.

This is for a project I am working on for my data-mining / statistics class. Any help would be appreciated and thanks so much in advance!

EDIT

I tried this:

I am trying to build a leaflet map where users can click once on a polygon to indicate that it has low importance, twice to indicate medium importance and three times to indicate high importance. I want the first time the polygon is clicked to turn yellow, the second time its clicked it changes to orange and the third time it changes to red.

I've found these two posts to change to red once the polygon is initially selected, and then once it is double clicked to remove it.

Changing styles when selecting and deselecting multiple polygons with Leaflet/Shiny

Select multiple items using map_click in leaflet, linked to selectizeInput() in shiny app (R)

A copy of the code mentioned above:

I am using Synth() package (see ftp://cran.r-project.org/pub/R/web/packages/Synth/Synth.pdf) in R.

This is a part of my data frame:

I'm trying to parallelize this piece of code that search for a max on a column. The problem is that the parallelize version runs slower than the serial

Probably the search of the pivot (max on a column) is slower due the syncrhonization on the maximum value and the index, right?

I've been off R for a few months, so that might have had some consequences.

I found this dataset on the internet. I treated it some, so I'll just dput() it here, but it originally came from https://ourworldindata.org/terrorism.

I am making choropleth maps with geopandas. I want to draw maps with two layers of borders: thinner ones for national states (geopandas default), and thicker ones for various economic communities. Is this doable in geopandas?

Here is an example:

In Python, I'm plotting a choropleth with some data for some countries in Africa:

I have a data source based on ISO3 country codes and I wish to visualize it using geopandas. The source I am using is based on world bank data and contains more countries than the target in GeoDataFrame.

I used the code below to achieve the visualization I was looking for. Some lines of codes correct missing abbreviations in the world layer (see https://github.com/geopandas/geopandas/issues/1041 for more info).

Now I have two problems.

1. (major) The code is not very elegant. Does someone know a more elegant was to import the data into the world GeoDataFrame?
2. (minor) I get a matplotlib warning I do not understand:

RuntimeWarning: invalid value encountered in less xa[xa < 0] = -1

Basic code to reproduce the problem is below. Sorry for the long arrays, but the data is necessary to reproduce the problem.

How do I create counts of how many observations were active between start and end years grouped by an id variable?

I have a df whose unit of analysis is organizations. These organizations are active for only certain periods of time (startyear, endyear) and in certain countries (acr). Some countries have more than one organization active at the same time. I would like to create a variable that reflects the number of contemporaneously active organizations for each country (for each observation the answer will be at least 1 and could be several dozen). For each observation's startyear, I believe I would need to go over my data frame by country and count how many times the startyear falls between other observations startyear/endyear. How to do it?

I have attempted to do this with dplyr (per Count how many fall in each group in R):