R is a powerful computation program, generally known as an environment for statistical computation and graphics. Nowadays this program is an essential tool for data scientists. The importance of R on research field is inevitable, when the datasets are complex and huge in number. The researchers have to be involved with collection, preparation, analysis, visualization, management, and preservation of large set of information. But unfortunately most of the researchers, especially biologists/ecologists are not have enough time to spend on coding,especially when they are in the middle of their research projects. But keep in mind the famous quote:
“Give me six hours to chop down a tree and I will spend the first four sharpening the axe.”
― Abraham Lincoln
Learning R for your research have same effect, it will sharpen your skills. From my own experience, I done my six month estimated data analysis within 6 hours. Most people thinks that R is just a program for statistical analysis. I must tell that, It is the bare minimum knowledge of R they have. In this post, I am trying to tell the possibilities of R. Whatever your research is, or what kind of data you are dealing with, R is the solution.
Automation of analysis
You can use R for automation of your analysis. This could be easy when you have lots of data with same structure and complexity. If it is not the same also you could do by make the code more smarter. You could plot all the datasets from a single folder(or multiple folders) within seconds. Write all the algorithms you could set, as the R only mounts specific libraries, the control flow swift with ease. Unlike other numerical computation programs it only take very little RAM(but depends on you data size).
Optimized tools for Numerical modeling & statistics
R is an optimized program, specifically for mathematical models. A wide varieties of tools from Comprehensive R Archive Network (CRAN) and inbuilt functions are available. All kinds of linear, non linear functions, time series analysis can be effectively done on R, but everything have to previously defined or you have to use user libraries(fortunately , you will get a lot from CRAN. Have a look on what R can do on models and analytics :
- Linear & Non linear modeling
- Multi level regression analysis
- Multivariate statistics.
- Time Series Analysis.
- Wavelet Analysis.
- Empirical Orthogonal Function
- Factor analysis(PCA/MCA/CCA
- .. .. . . . …………………………….
Automate image editing ,processing and shape analysis
Most researchers depends on various image editing softwares to edit their images for their day today research life and everyone have their own favorite lists to do this. I can’t tell you that you can do all the image editing in R like in Photoshop, but you can do more than that. For example if you want to change the background of a picture, what will u do?. you could simply do it on Photoshop or any other contemporary image editing softwares. It is surely a simple job, but how is it when you have 100 more, or 1000 images. There R is …!!!!. You could batch process, automate the editing process with minimum effort.Another essential feature you could make use of R is feature extraction and statistics from imageries. It is a trending topic as it is widely used in computer vision, machine learning and AI. Most used shape analysis methods such as elliptical Fourier descriptor(EFD), Thin plate splines(TPS) can be easily executed.
Fine tune your graphics
R is famous for its graphics. From my perspective I like more on its fine tuning capability in the graphical display, in which you can do what exactly your outlook from the data. You don’t have any limit, or constraints for a plot you imagine. But I am sure, for the first you get suffer, which is obvious. As far as I know, there are three graphic systems : a) The base one b) ggplot c) Lattice. I will tell about more on these things on my later posts
Geo-spatial Analysis and Remote sensing with ease.
Most ecologists/ Earth science researchers need to address there topic’s spatial relevance and dynamics. Now a days this is somewhat easy by remote sensing , A lot of satellites are revolving around to do this work and are mostly free to access. The dark side is ..you have to download and process tons of MBs/GBs of huge data, that hang your PC. This is happened because you need to download the entire world map(In some Live access servers(LAS), things are different ), mostly when you want level 1 datasets. The wrong thing is ; we are very happy with GUI and do browsing through the map, this take lot of RAM and again make your PC hot. By using R you can skip the previewing of data, but you can conjure all the variables you need. You can crop with thousands of bands with in seconds. Here some of the analysis I done recently, but you do many more.
Spatio-temporal Data Analysis.
Spatial statistics, including spatial auto correlation and spatial regression.
Bathymetry and depth contour analysis.
Sampling strategy aids.
- Pixel extraction
- band stacking.
- spatial modeling and simulation.
- empirical orthogonal teleconnections.
Get the data direct to your console without any web interface
This is another advantage I feel like awesome. You could connect the online data direct to R(but you have to know the url). Usually we first download the data using a browser and import to the software of interest, start doing the processing game. If you are in R you can skip the step, and connect directly to the data. Before doing this, you have to aware about your PC’s capabilities and size of the data you are going to connect.
Do all your documentation in R.
when I got grip on R, I take an oath that I could do all my documentations in R. At the beginning , I thought I just fancied because I like R, which made a lot of leisure time for me. However, the documentation power of R is beyond I could imagine. To be frank, I can’t explore much on this documentation for my paper writing. I done some pdf extractions, and had a try on presentations. It feels great!!. Here the list I found that you can do on documentations.
- R markdown and KnitR for writing and export to docs
- R presentations are usually html or pdf(But there are ways to convert to ppt).
- Make beautiful animations for your presentations.
- RefManageR- reference manager/citation
- Pdf reading and extarction
- All the document loving formats can easily edit(.csv,.doc,.html, .nc,.text……and many more)
- Latex/bibitex editing
- Finally make your own database using RSQLite, An awesome method to make your research in order.
Connect to the real world.
All the above features explained are the capabilities of R for data analysis. This is something different. You could connect the R with open source hardwares such as Arduino(An open source microcontroller platform for hobbyists ), EV3 linux distributions/Raspberry PI and you can experiment as your wish. This means you can take data directly from an instrument(device/robot/sensor) to process. This would make a big leap in amatuer robotics, especially on machine learning. I want to tell this because you should know the fact that R is a growing platform. As a researcher we should not limit our experiments for instruments, which are available only on the market.
Click the pics for more details
Everything under one umbrella
The most difficult and the most interesting part of research is you have to deal with a lot of different topics, which have to connect logically(according to your knowledge and experience). As every research ends in data, (and starts from data) you have to run for extremely different approaches to extract vital information. As an ordinary researcher, you have to depend on hundreds of softwares like I above mention, for image editing Photoshop, for Mapping ArcGIS, for PCA primer………. . . . . . ……………………………….. this will not ends. The R makes all these jobs under one umbrella, This makes R more special.You could mount an image editing package simultaneously with a statistical package. The best part is R could literally convert any dataset to numbers, that you can analyse.
Make a domain of yourself
R program is not a proprietary program, it is a free one(open source), which means it respects user’s rights. You could develop, engage, use, reuse and modify for your needs and share with your friends for a try. The learning gap between a programmer and developer in R is seemed to be narrow. I am not a programmer, but I develop a couple of packages(Not yet published).In R world everyone have a space to grow independent of your subject area. I hope this post make an overview of what you can do on R. But I am sure that I didn’t explain everything about R. I promise you that my upcoming posts will explain things in detail.