-
Notifications
You must be signed in to change notification settings - Fork 31
FireData Connecting R to Google Firebase
Project Title: FireData: Connecting R to Google Firebase short title: Creating a Firebase API
Persistent Data Storage and transfer with R can be complicated and time consuming. While there are great packages to connect R to DBs like mySQL/sqlLite/postgreSQL those databases itself need to be hosted and maintained. To implement a DB based login system in Shiny now probably takes 2 weeks of work to get it done in a secure way, and a IoT project for data collection requires a 24/7 running server.
The Google team took that common problem (not only of R) and created Firebase. Firebase is a online database that is not only intuitive to integrate into Apps & Web applications, but can also be used to build user registration/login systems with persistent data storage (thinking about shiny) or to transfer data between IoT devices via Firebase directly to a R environment. Writing a intuitive R to Firebase connector enables R to be used in new "cross platform" realtime applications. Thats why we strongly believe the "FireData" package will improve the R open source environment.
The R universe offers with the httr package to apply PUT/POST commands to servers (including Firebase). In combination with a JSON library like jsonlite it is possible to send & retrieve data from Firebase.
Google Firebase on the contrary is more sophisticated than simple PUT/POST requests and allows JWT tokenizations, registrations & logins and even file uploads. Firebase brings all functionality and ease to be the to-go Database for R IoT or Shiny related projects.
With the experience of GSOC 2016 we will start with a larger team (other open source developers signaled it would help to improve their own DB experience working with R and Shiny) to get the FireData package CRAN ready before the September deadline.
- Now - April : Adding prove of concepts of the Firebase integration into Shiny and IoT applications.
- April - June : Writing and implementing (including unit test) the various firebase connecting functions
- June - July : Adding the s4 classes/functions to the FireData package & writing Vignettes/Documentation
- July - August : Finishing up the FireData package with focus on OS compatibility testing and identifiying and fixing weakpoints of the package.
R is one of the strongest players in data science. The aim is to connect its strength in data analysis with the actual data. By making it easier to directly connect R to the inputs of IoT devices or other applications such as websites/shiny it will open new possibilities for open source developers around the world.
The final product, the FireData package, will connect two great environments with each other. When we succeed to make it stable & intuitive alot of projects will be build ontop of it.
Daily submits to the github repo. The project will be started in a test-driven fashion by writing "testhat" tests for every method implemented. For the Firebase testing we will accordingly create a test database (firebase allows to upload xml database mirrors which correspond to real db environments).
- A tutorial on Shiny & Firebase Logins written by the student https://r.codes/ItemPreview/-KaLAwn-zqMgARFUnxpo
- proof that student can create R packages with roxygen2 and devtools: https://github.com/Kohze/Discovr
(the R server side Firebase proof will be added soon)
Samuel Schmidt, Ph.D . (has GSOC Mentor experience). | schmidt@glycosearch.com
Lindolfo Pedraza, MA. (Database and API specialist) | lindolfo.pedraza@gmail.com
Robin Kohze | https://github.com/Kohze | Master student at the Radboud University | Project leader of https://R.Codes | robin@kohze.com | GSOC student of 2016 for the discovr package
Student affiliation:
- Institution: Radboud University Nijmegen
- Program: Medical Epigenomics
- Stage of completion: Master | 1st year
- Contact to verify: Can provide official documents