Hacking the language barrier for a billion-plus people

By Sindhu Kashyaap
November 11, 2015, Updated on : Thu Sep 05 2019 07:24:30 GMT+0000
Hacking the language barrier for a billion-plus people
  • +0
    Clap Icon
Share on
  • +0
    Clap Icon
Share on
Share on

What if you had the chance to solve real-world problems, and change the lives of over a billion people? This was the challenge posed at the YourStory TechSparks 2015 Hackathon. The 16-hour long hackathon, which focused on language and commerce tracks kicked off on the 29th of October. Our partners TeamChat, Dailyhunt, Reverie, Loginext, and Exotel exposed APIs and technologies to enable participants to develop applications and drive India- specific innovations.

Yourstory- TechSparks- Hackathon

The tracks were:

Track One: Language-focussed:

Developers had to use APIs from Reverie and Dailyhunt to marry content, community, and communication to develop mobile-first apps for Indian local language users; the challenge was to come up with out-of-the-box ideas with scalable solutions.

Track Two:

Commerce focused – Developers will use the SDK of our partners Teamchat (Java based) and Loginext to come up with business solutions.

Track 2.1

What are the real coordinates of your customer?

Loginext provided sample data for addresses and verified coordinates (latitude and longitudes).

BASIC: New addresses to be geocoded (converting addresses to latitude-longitude) using different data sources e.g. Google Maps API, OSM, MapMyIndia API, and PitneyBowes, etc.

– Use Machine Learning to find out the most accurate latitude-longitude on basis of the previously geocoded or verified addresses.

2. ADVANCED: Smartly jump between data sources to get the accurate traffic information or to calculate accurate ETA.

*Think of ways to integrate local languages here too.

Track 2.2

Using smart messaging paradigm, develop vertical specific client solutions. Example, case studies available on: - http://www.teamchat.com/en/case-studies/

Using smart messaging paradigm, integrate Teamchat with external applications like Jira, and Facebook etc. Using smart messaging paradigm, create utility bots like weather updates, recipe, and jokes etc. Examples available on:- http://www.teamchat.com/en/bot-store/

*Think of integrating local languages here too.

Three teams emerged the winners with their innovations:

Team Builders - First place

A team that came together only on the day of the hackathon – Team Builders created an app called Dragoman, a smart personal translation assistant app, that works across all apps. The motivation behind the development of the app was crossing language barriers. When an average user on WhatsApp or Hike is part of a group, and a person texts in his or her native language, only those people who know the language can respond, while others are left out of the conversation. The team initially thought of making it a multilingual chat app, but then realised that even if they made a useful alternative app people would still be reluctant to move from WhatsApp or other apps.

Dragoman has a floating UI similar to Facebook's chat heads, which is always available on the screen (non-intrusively). Whenever a person sends a message on any messaging app in his/her native language, all you have to do is long click on the text and it auto-magically translates the text into a language you are comfortable with.

In the backend, the selected text is sent to Reverie's transliteration API, which uses Google Translate to shows the translated text on the floating UI bubble.

Team Kaos - Second Place

The team worked to address a problem non-English speaking users faced with a majority of apps, where the content and navigation are in English. They also aimed to address user- generated content/messages. E.g. If a non-English speaker wanted to book tickets for a movie, he or she would not be able to navigate an English-only booking app. This gave birth to the app 'Akcessible'

Team Ekta- Third place

While there are a large number of excellent tutorials and Massive Open Online Courses (MOOC) available online, regional language speakers in India are unable to utilise them as the content is largely in English. While Google Translate provides an easy option to translate websites to Indian languages, it can fall flat when it comes to tutorial sites (e.g. Khan Academy) or MOOC websites, as the translation can be inaccurate, which makes the content unusable.

The team developed a JavaScript framework, which makes use of the excellent, easy-to-use REST API from REVERIE to translate the content to the user's preferred local language. As there is no server-side processing required, the advantage is that existing sites need not be modified in any way.

The three winning teams not only made their mark with their brilliant innovations, they also helped realise the TechSparks 2015 objective: Technology that can transform the lives of India’s billion-plus people in ways that nobody ever imagined.