readme.rmd

---
title: "Readme"
author: "Marc Trunjer Kusk Nielsen"
date: "11 maj 2016"
output: md_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
library(MadsR)
library(dplyr)
library(ggplot2)
```

MadsR - let's get to the fun part.

MadsR is a package designed for my own uses with data from the danish microbiological system MADS.

Installation
----------------
```{r, eval=FALSE}
devtools::install_github("marcmtk/MadsR")
```

Functions implemented
-----------------
This package provides 6 functions to assist with epidemiological analyses of MADS data and 2 functions to generate MADS like data for testing purposes.

* `read_mads` - reads a csv file, applies appropriate types to columns, adds time columns and splits name of sender into useful categories.
* `filter_cases` - easy filtering of first episode in a timewindow.
* `since_last` - calculates the time since a new positive sample was last seen.
* `tally_by_department` - calculates positive cases by department, provides functions to exclude observations based on different windows, see examples
* `slplot` - a "Since last" plot
* `tally_map` - a heatmap of tally data
* `generate_MADS_like_data` - as the name implies 
* `generate_res` - generates resistance patterns

Examples
---------------
Consider the dataset provided in analyser-like.csv with read_csv vs read_mads

```{r, warning=FALSE}
df <- read.csv("./extdata/analyser-like.csv")
str(df)
head(df)
df <- read_mads("./extdata/analyser-like.csv", "analyser")
str(df)
head(df)
```

Let's filter some cases and look at the results
```{r}
cases <- filter_cases(df, result=="Positiv", min.days.to.new.episode=14)
table(cases$hosp_afd, cases$episode)
table(cases$sl)
```

Now let's look at a since last plot, after all there may be an epidemic out there!
```{r, warning=FALSE}
since_last(df) %>% slplot()
since_last(df) %>% filter(hosp_afd=="S V") %>% slplot()
filter(df, hosp_afd == "S V") %>% since_last() %>% slplot()
```

Note the difference between plot 2 and plot 3. It is very important that time since last positive case is computed *after* relevant filtering.

Last useful functions, tallying by department and heatmapping the results:1
```{r, fig.height=4, fig.width=12}
tbd <- tally_by_department(df, "patient", result == "Positiv")
tally_map(tbd)
filter(tbd, hosp_afd != "AP ") %>% tally_map
```


Session info
-------------
```{r}
sessionInfo()
```