Brief Introduction

The Indic Language Visualization Tool helps to translate words (supported by multilingual word vectors) and visualize them in a 2 dimensional space (using t-SNE on the word vectors).

Languages supported : English, Hindi, Bengali, Marathi, Gujrati and Tamil

Dependencies

• Python (version 2.7)

• For proper visualization, you might need support for Indic languages.

Libraries

scipy, sklearn, cgitb, cgi, numpy, pandas, matplotlib, mpldatacursor, Tkinter, PIL

Data Download

Download the input data (word vectors & word similarity list) here

Unzip the file before executing the code

Usage

1. Execute the python code

2. Select the mode

• Mode 1 - Enter the number of words : In this mode you need to provide the number of words per language you want in the translation

• Mode 2 - Show top similar words

3. Input a query word

• Use the suggestion list and select a word

• Or, Type the word and click on 'Find !'

Output :

For Mode 1

For Mode 2

• Window 1 : Translated Words

• Window 2 : Visualization

Acknowledgements :

We would like to thank Ashim Gupta and Ravi Shankar for their help.