Data Exploration and Machine Learning app using R Shiny

shiny

#1

Team,

I am sharing the app related to Data Exploration and machine learning. Features includes:Data Exploration, Running Models, Finding important Variable, Comparison of the Model and Choosing the best Model to predict.

In the next version I will include features to deal with missing values, outliers and transformations. Hope it helps.

Keep using and let me know your suggestion and issues/bugs you are facing.
H
http://suryaprakash.shinyapps.io/Data_Exploration


#2

Hello @Surya1987 This is great. Would you mind sharing the code?


#3

This is awesome. Thanks a lot for sharing the app. Have been always thinking about developing a similar app and now you just saved me. However, i have few suggestions:

  • can you add an option to save the trained model for a later use
  • it would be great if the user can also upload a custom machine learning model

Are you planning to open source the app?


#4

Yes I will add those features which you suggested. Currently it is open to everyone. Just I am trying to create an which will be helpful to everyone. Hoping for the best. Saving I am just thinking how to do. Write.csv does not work on server. Let me know if you have any idea. Have u explored the app? Let me know if you face any bugs.


#5

I did a quick try on a classification problem but it seems not to respond when I try to train. The dataset is fairly small (220KB) 90000 rows x 12 columns.

As for saving the trained model, I usually save them as pickle in python but am not sure which R version does the same (maybe dput and dget?). The model shouldn’t be saved on the server but the app would offer the flexibility to save it on a local directory.

Another suggestion I have is to generate tables as tables instead of text. For example, the correlation is displayed as text similar to how R generate correlation. If you could have this correlation matrix and feed it to Shiny so it output a structured table would be better.

This is definitely a good project and if the code is also open-sourced, am sure there will be more people adding more features.


#6

Just make sure that the dataset you upload should not have missing values


#7

I have noted down your suggestions. I will update/add one by one. Currently working on code related to dealing with missing values, outliers and transformations


#8

Hi Surya,

Can you please share the code?


#9

Hi @Surya1987 can you please send me the code for that, i am also working on same problem, please send it to my mail id peddollakishore@gmail.com.

Thanks in advance


#10

@Surya1987 could you please share the code. my id is dippy.018@gmail.com


#11

Hi Team,

Here with I am sharing the code for the code. Due to some project work I was not able to improve it further. Kindly improve the app with adding new features which will help everyone.

There could be loop holes in the current project which can be fixed and optimized the code as well.

[[https://gist.github.com/Surya1987/d0f91c892e8f0bbfb225d21d10e46e89]]


#12

Hi Surya,
This is Great one.
I have created one app. but, I want to know how to share the app. Can u please tell me how to share the app into link.


#13

please share the information on his email id:- jdivyabharathi7496@gmail.com for my above queries…


#14

For that you need to have your own server. Otherwise you can register in Shinyapps.io and publish your app for free which will have basic features. You can follow the steps mentioned in below link.

Pricing will be as per your need. For details vist
https://shiny.rstudio.com/deploy/
http://docs.rstudio.com/shinyapps.io/


#15

@Surya1987 Thank you for your Reply…


#16

@Surya1987 The App was really wonderful Can you Please share the code to maloojinesh@gmail.com…I am also working on somewhat similar problem