sepmein / snt

Analysis code for malaria

snt

Codes for malaria analysis

Maintainer: Spencer [email protected]

Get started

The code is written in R, to access the package, please use devtools to install.

Install the devtools library if not.

install.packages("devtools")

devtools::install_github("sepmein/snt")

Functions

Get multiple files

Raster file extraction

snt's Issues

Template for projects

When creating a new projects, it could be quite normal to copy paste codes from other projects.

These codes include comments to refine R code structure, codes to extract rainfall, population or other raster data.

Code templates could let us get started much faster.

Add Color Scheme in the project

When generating maps for SNT analysis, maps are often needed to be presented. We often need to present a combination of interventions through colors.

In R, there are a few packages and function allow us to generate colors.

Add a central configuration file and configuration function

Add a snt::config function that will be allowed to store the default location.

Show provinces in the map

ADM1 boundary currently do not show in the map. Need to load from shapefile. If no shapefile provided by country, need to calculate from ADM2 districts.
@sepmein

Nigeria RIA analysis

Correlation test of two sets of databases

When comparing the set of data frame, a correlation test was needed to check the correlations between two datasets.

There are a few algorithm built in R or available from other packages. I need to

1. identifies the nice packages
2. pick some data columns for example from routine database, e.g. conf, test...
3. created plot function

Create a line graph with ggplot2

Create a universal line graph that could be used to display various information such as rainfall, prevalence, mortality.

• Use ggplot2 facets to display multiple graphs.
• Try to add LCI and UCI with shades into the graphs.

Somalian Adjusted Incidence

To create an RMD to automatically generate a report of the column names of each input files

When doing data management, we often need to review the data columns to get insight of the raw data structure. It would be really nice if we could have a document automatically generated reports could reflect on the data structures.

Update rainfall importation codes

After the extraction of rainfall raster file, I need to manually extract year and month from the filename as follows.

resource_rainfall_adm2$data |>
    mutate(
        year = stringr::str_extract(file, "\\d{4}"),
        month = stringr::str_extract(file, "\\d{2}(?=.tif)")
    ) 

So I need to manually do it every time I need to extract file.

The better way is to incorporate this codes inside the Rainfall resource class.

Add a summation method for raster file

When calculating pop in a district, snt should be able to sum all the numbers in the area.

Enhancement: Refine structure with usethis::use_r()

There are a large chunk of code lack of management in main.r, I need to reconstruct them into separate files.

Parallel execution for raster resource

In the main.R RasterResource load_single_folder function, I'm now using normal for function for the loop, it is slow and not using cpu. To speed up execution, should introduce parallel execution.

The raster resource didn't load correctly

Raster resource loaded multiple times

Add String Fuzzy Matching function

When importing data, some of the data columns often needed to be check duplicates. Data in the columns often not exactly matches.

A fuzzy match function need to be developed for faster duplication detection.

A very promising function was described here in the blog: https://blog.devgenius.io/approximate-string-matching-in-r-using-jaro-winkler-similarity-a93436ecf38f

Import Routine Data

Separate cleaning, importation, analysation, plot steps

Data process could be separated into several steps.