pollster | A minimal async executor that lets you block on a future | Reactive Programming library

I have a frequency table aggregated from 800 millions of records and am wondering if I can use a package to calculate 1st order transition matrix from the frequency table, which is not symmetric because some state just never happened again. A sample of the frequency table is:

I can't find why is the error happening.

The dataframe looks like:

I have created a user function to generate cross tables for my categorical survey data. For example, I have written a function to find Q50_1 by gender. It looks great.

But now, would like to add a column, with the corresponding SE for each row, but I cannot figure out a solution. Here is my desired output (the SE's are fake in this example).

Here is my data:

I'd like to feed two vectors into my custom function with purrr:map2. Unfortunately, when I do this, I get the error Error: Problem with `filter()` input `..2`. x Input `..2` must be of size 93 or 1, not size 9. ℹ Input `..2` is `!is.na(y)`.

How can I get my custom function to use arguments as data and run successfully with map2?

This is a MWE:

I am trying to find the weighted cross tabulations of several columns. My actual data has over 500 columns so I am trying to automate the process as much as possible.

This is a snippet of my data:

I have a system where I'm using the consensus opinion of weighted votes to predict a binary outcome.

Since elections are topical we can use it as an example. Say I do an analysis on various pollsters over the years and assign them a weighted vote based on how accurate they were. Pollster Skyler ends up with a vote weight of 3 and Pollster Oakely was twice as accurate and ends up with a vote weight of 6. In CGP Grey fashion (a youtuber who sometimes talks about election mechanics) Skyler predicts that the Tiger candidate will win the open city council seat and Oakley predicts that the SnowLepoard candidate will win it. In this example, the projected winner, based on total weighted votes, would be SnowLepoard with a majority weighted vote percent of (6/9) 0.66667. The actual outcome of any given election would of course vary, but generally, if the weights are good, we'd expect the win probability of a given candidate to increases as the majority weighted vote percent increases and, likewise, if the majority percent is 50% for the election to be a real tossup.

With that example in mind, I'm trying to do a logistic regression on a dataset where the only independent variable is the majority weighted vote percent. I've tried a couple of different methods and while both generally agreed that win probability increases as majority weighted vote percent increases, they disagree as to how much and neither really respects the idea that a 50/50 majority weighted vote percent indicates a true 50/50 probability.

The blue line is the logistic regression done through seaborn and the green dots are the logistic regression done through sklearn. Code below. I dont think that the underlying mechanics of a logistic regression change from one library to another so clearly if they're producing different outputs for the same input, my setup is wrong.

• Why are these two libraries producing different regressions?
• How can I force the regression, for either library, to treat a weighted vote majority of 0.5 as a 50% win probability? I can probably just fill in a mass of dummy data to force the conclusion but I feel like theres got to be a more elegant way.

I am following along Traversy Media's Django tutorial for the poll app and when I ran my code near the end I got the error stated in the title all of the sudden: why is detail not found?

views.py