Hult Prize 2019: The Global Unemployment Challenge

Posted by

This is the visualization presented by the United Nations on their 8th Sustainable Development Goal, Youth Unemployment. The Darker regions indicate high levels of youth unemployment. Does this make sense to you? Some are quite surprising.

Overview of the UN’s SDG 8 – Youth Unemployment

I recently competed in the 2019 Hult Prize Challenge, and presented a concept to allow native speakers of endangered languages the capability to produce and share online content centered around Entrepreneurship. For the investment of USD1M, the start-up would compensate with USD9/hour within a 10 year period and have an ROI of approx. 38%.

The Challenge

The UN’s Sustainable Development Initiative has a tool that allows you to track the progress of each SDG. We see that the highest degree of youth unemployment occurs regions shaded red and black, indicated potential markets of application.

From here, the team took a Red Team, Blue Team approach to diversify our innovation potential. Each team member spent 2 months accumulating their resources and focusing on their own implementation, and presented their recommendation on the third. I chose to approach the problem by understanding resources they currently (mostly) have; digital technology.

Using PythonPandas and matplotlib I put together and wrangled unemployment data from the International Labor Organization (ILO), The United Nations (UN) Open Databases, The UN’s ITU and the World Bank. The Python code for this can be found on my Github. Here are the findings.

Findings from wrangled data using Python, Pandas and matplotlib

The wrangled data revealed a surprising phenomena; unemployed youth that were connected to the internet had steadily increased over the last 4 years, and expected to continue. Furthermore, and this surprised me,  80% of the available information on the internet was only available in 10 languages. With a statistic like this, I’m not surprised of what most Sociologists refer to as a “Digital Divide“.

Finally, given that Google expects 30% of its search volume to come from sources other then a screen, it made sense to tailor a solution that around a Speech-To-Text application.

The Proposal

Byblos – A Speech-To-Text Solution to generate content around Entrepreneurship

After many rounds of feedback from our mentors, and discussions within the group, we settled on a Speech-To-Text solution, Byblos. The proposed value chain was as follows,

  1. The user would generate content in their native language by converting the data from their voice into text.
  2. This data would be housed on our servers, using SalesForce as our primary solutions provider, and sold to potential customers that need it. This would include Linguistic Museums, Open NLP Tools & Repositories, Linguistic Anthropology Societies, Private Companies and Large Technology companies such as Google, Amazon, IBM and Microsoft.
  3. These companies would have access to this platform, with the opportunity to generate paid advertising in various native languages.

The Conclusion

This was certainly a fun project! I pushed my Python Data Wrangling and Analysis skills beyond my capabilities, as well as learned that many people don’t have the same opportunities as I do. We weren’t selected for the next round, and learned a lot more about myself and my capability to network and exercise a healthy growth mindset.

How do you think this solution could have been improved? Looking forward to hearing your thoughts!

Leave a Reply