A web application for AUTO ML

Building a web application is one of the most important skills in Data Science projects. This post shows an app that is designed for building a model directly on a server with XGBoost, GLM, Deep Learning, Random Forest and GBM.

Turgut Abdullayev https://github.com/henry090 (AccessBank Azerbaijan)https://www.accessbank.az/en/
01-22-2020

Table of Contents


Intro

Building a powerful web application in R requires knowledge of R, Shiny framework, and a bit of CSS, HTML. R skills are enough for building the structure of the app, but it is not enough to make an app beautiful and much more user-friendly.

In this post, we would like to present an application which is very helpful for companies to build a model without a knowledge/experience in data science.

Authentication

Authentication allows to limit an access to your application. Here you can even set limited time in access to your application. So, it is very useful to include such feature to your application and prevent extensive traffic to your server.

What you need to do

Upload your data via “Browse”. This section will quickly read your data but keep in mind that it is better to upload file in a csv format because excel files can be very slow in reading process. As soon as one uploads a csv file, the app will immediately show the summary of the dataset. Summary covers very useful information about:

Train different models

In order to begin the training process one should choose the column which needs to be predicted. But there is also a menu for choosing an algorithm and at the same time to point out whether a user wants to stack the models or not.

Result

For now the system itself decides the approximate time for training a model(s). The main criteria is the size of dataset, particularly the row and column numbers of the data.

This page covers a lot of valuable information:

Further improvements

In the future we plan to add the number of models that user wants to build. Furthermore, an ability to get an automatic report of the results of the modeling would be very useful and effective for non-technicians. We plan to knit a report in the form of HTML, Word, PDF, and Power Point.

If you have suggestions or would like to test the application, please let us know by writing a comment below.

Corrections

If you see mistakes or want to suggest changes, please create an issue on the source repository.

Reuse

Text and figures are licensed under Creative Commons Attribution CC BY 4.0. Source code is available at https://github.com/henry090/dataexperts_blog, unless otherwise noted. The figures that have been reused from other sources don't fall under this license and can be recognized by a note in their caption: "Figure from ...".

Citation

For attribution, please cite this work as

Abdullayev (2020, Jan. 22). Data Experts: A web application for AUTO ML. Retrieved from http://dataexperts.tech/posts/2020-01-22-automl/

BibTeX citation

@misc{abdullayev2020a,
  author = {Abdullayev, Turgut},
  title = {Data Experts: A web application for AUTO ML},
  url = {http://dataexperts.tech/posts/2020-01-22-automl/},
  year = {2020}
}