firstBinMedianBug.Rmd

---
title: "Testing bugfix in version 0.2.2"
author: "David J. Hunter"
date: "March 11, 2020"
output: html_document
---

# Testing the inverse CDF

The CDF is in blue, and its inverse is purple. Below, we use `sb_sample` to draw a sample of 200000 points, and then compare that (in black) to the `splinePDF` (in red).

## Cook County data

```{r}
library(binsmooth)
library(ggplot2)
binedges <- c(10000,15000,20000,25000,30000,35000,40000,45000,
50000,60000,75000,100000,125000,150000,200000,NA)
bincounts <- c(157532,97369,102673,100888,90835,94191,87688,90481,
79816,153581,195430,240948,155139,94527,92166,103217)
splinefit <- splinebins(binedges, bincounts, 76091)
ggplot(data.frame(x = c(0,splinefit$E)), aes(x)) + 
  stat_function(fun = splinefit$splineCDF, n = 1000, color="blue")
ggplot(data.frame(x = c(0,1)), aes(x)) + 
  stat_function(fun = splinefit$splineInvCDF, n = 1000, color="purple")
df <- data.frame(x = sb_sample(splinefit,200000))
ggplot(df) + 
  geom_density(aes(x=x)) + 
  stat_function(fun = splinefit$splinePDF, color="red",n=1000) + 
  xlim(0,splinefit$E)
```

## What happens if the median is in the first bin?

In Version 0.2.1, the following code block generates `Error in if (fitWarn) warning("Inaccurate fit detected. Proceed with caution.\n") : argument is of length zero`. Fixed in Version 0.2.2.

```{r}
binedges <- c(10000,15000,20000,25000,30000,35000,40000,45000,
50000,60000,75000,100000,125000,150000,200000,NA)
bincounts <- c(2000,100,100,100,100,100,100,100,
100,100,100,100,100,100,100,100)
splinefit <- splinebins(binedges, bincounts)
samp <- sb_sample(splinefit,200000)
df <- data.frame(x = samp)
ggplot(df) + 
  geom_density(aes(x=x)) + 
  stat_function(fun = splinefit$splinePDF, color="red",n=1000) + 
  xlim(0,splinefit$E)
median(samp)
mean(samp)
sb_percentiles(splinefit)
```