Loading
O17-2020: Team N "One"

O17-2020: Team N "One"

by gowtham.s | updated April 21, 2020
based on O17-2020: Team XX

Tracking Sustainability through History

0

CHALLENGE 5 - TRACKING SUSTAINABILITY THROUGH HISTORICAL TRENDS:

Project Description:

The challenge is to track sustainability through historical trends. These historical trends can be found from the digitized League of Nation's documents in the UN Library. This is very important to solve because in that millions of raw texts called data there will be some useful information that is hiding within it. For instance, those Information can help us provide some insights on how to tackle a similar problem that has happened in the past.

I propose an AI model that extracts important information from the vast amount of data by building a knowledge graph and also creating a platform so other people can easily use it.

The approach is better because we create not only a decision making system but also a general platform where everyone can contribute and to learn about the history of the world.

Team Introduction:

I am Gowtham, a self taught budding Data Scientist, I code well in Python and I am aware of AI concepts. I am in a Single Team. I am doing my last year B.Tech Computer Science and Engineering from Sikkim Manipal Institute in India.

March 30, 2020 at 10:21 AM
Created by amudha
Edited by gowtham.s
Comments (2)
Make sure to fill in the required info before the next session. The sooner you do it before the session, the longer we have to give you feedback to improve your project.
3 months ago
Hey Gowtham! I hope you are ok and look forward to working with you. Please let us know how things are going.
Challenge 5 was addressed in the last session so it might be nice for you to watch the recording of it. :)
3 months ago

Please sign in to leave comments

Describe what is the need of this project?

I need DATA.

That Data is already gathered by the efforts of our previous Summer School members which is shown in the image above, the crowd-sourced digitization of League of Nation's Archives. Some examples of the type of Data is also present here https://libraryresources.unog.ch/c.php?g=462663&p=3163203 .

This project also needs AI algorithms to extract the information required.

March 30, 2020 at 10:21 AM
Created by amudha
Edited by gowtham.s
Comments (0)

Please sign in to leave comments

Describe who is affected?

In a single world, EVERYONE.

Information should be available in a global and a friendly manner so any people can use it, explore it and that could affect everyone one in the world.

For example, the above picture shows the pandemic of Influenza which was part of League of Nation's Archives and that actually gives us a heads up and certain measures to be taken against such outbreaks, Information like these should come out for the greater good.

March 30, 2020 at 10:21 AM
Created by amudha
Edited by gowtham.s
Comments (0)

Please sign in to leave comments

What are the causes?

The major cause of this problem is the unavailability of systems that can carry out the extraction of important information from the millions of structured and unstructured data.

I also feel that there are millions or at-least thousands of excellent Data Scientists out there in the open source world who would come forward and do an Exploratory Data Analysis but there isn't any friendly API's that grants access to the millions of freely available data.

March 30, 2020 at 10:21 AM
Created by amudha
Edited by gowtham.s
Comments (0)

Please sign in to leave comments

`What is the evidence? Also who can you interview? What can you find out? What experiment can you run?

The above picture shows that Researchers at Bristol used 150 years of Britain newspaper articles ranging from the year 1800 to 1950 to extract certain insights within those papers and that shows the model I proposed is doable.

Enterprises around the world uses AI to create Knowledge Graphs to make certain business decisions and to track their financial growth which more or less leans towards our goals.

March 30, 2020 at 10:21 AM
Created by amudha
Edited by gowtham.s
Comments (0)

Please sign in to leave comments

Start documenting your thoughts and ideas

March 30, 2020 at 10:21 AM
Created by amudha
Comments (0)

Please sign in to leave comments

What is The Big Idea? What is the value proposition?

The data that when it can be extracted and understood can be used to predict any future occurrence of similar kind.

For example: The Fall of a Government's economy might correlate to any other country's conditions years ago, and having that knowledge or knowing that a similar event might occur can get us act fast.

March 30, 2020 at 10:21 AM
Created by amudha
Edited by gowtham.s
Comments (0)

Please sign in to leave comments

What is the mechanism of beneficial change?

1. People around the world can access our archive much easily and help us find any useful information that we have missed. 'Thousand eyes are better than two'.

2. People can learn about the history, culture of many nations and can get to know about documents of significant importance, example: Alfred Nobel's letter about the Noble Prize.

3. We can take preemptive actions on those events in the present that has a very strong correlation to any unfortunate events that would have happened in the past.

March 30, 2020 at 10:21 AM
Created by amudha
Edited by gowtham.s
Comments (0)

Please sign in to leave comments

What are the key metrics?

1. When a perfect Knowledge graph is created with which anyone access data easily.

2. To find any present world data correlating with the past data.

March 30, 2020 at 10:21 AM
Created by amudha
Edited by gowtham.s
Comments (0)

Please sign in to leave comments

Who is most likely to be supportive?

Open Source coders and Researchers out there who wants to use the data and find some interesting insights.

The UN library, Researchers who are part of this initiative.

March 30, 2020 at 10:21 AM
Created by amudha
Edited by gowtham.s
Comments (0)

Please sign in to leave comments

Key foes? Who is most likely to oppose?

1. Machine Learning Algorithms that would throw lot of errors but would eventually join our way.

2. Hardware capabilities to run the algorithms if not done in cloud.

March 30, 2020 at 10:21 AM
Created by amudha
Edited by gowtham.s
Comments (0)

Please sign in to leave comments

What is the user experience?

1. Any user without any coding capabilities can open our UN library website and if any keywords are typed for example, pollution: Then all the documents containing the keyword pops up. Basically Knowledge Graph gives the ability to be used as a Search Engine.

2. Users with coding experience who want to perform any analysis can use our API to fetch the data directly into their code and perform any analysis.

3. In the back end, we can continue with any research to try and predict any future outcomes on any specific topics such health, finance, pollution etc.

March 30, 2020 at 10:21 AM
Created by amudha
Edited by gowtham.s
Comments (0)

Please sign in to leave comments

Who has to do what. to make it happen?

1. UN Library:

        - Provide access to the infrastructure to build the system.

2. Team N:

        - Will build the infrastructure.

 

March 30, 2020 at 10:21 AM
Created by amudha
Edited by gowtham.s
Comments (0)

Please sign in to leave comments

Who are the key partners to execute? Key partners to help others evaluate your value preposition?

1. UN Library:

     - To execute the access of documents for analysis.

2. Team-N:
     - To execute the proposed model.

Partners to Evaluate:

1. Open 17 staff team.

2. AI experts.

3. Open Source Coders.

4. Users accessing the archive.

5. Library.

March 30, 2020 at 10:21 AM
Created by amudha
Edited by gowtham.s
Comments (0)

Please sign in to leave comments

What are the precipitating events?

1. We have digitized our archive.

2. We have loads of data that can be put to good use.

3. Rise of AI in current world enabled us this opportunity.

March 30, 2020 at 10:21 AM
Created by amudha
Edited by gowtham.s
Comments (0)

Please sign in to leave comments

who else is in the field?

March 30, 2020 at 10:21 AM
Created by amudha
Comments (0)

Please sign in to leave comments

What's wrong? Missing? Not working?

The approach is better because we create not only a decision making system but also a general platform where everyone can contribute and to learn about the history of the world.

March 30, 2020 at 10:21 AM
Created by amudha
Edited by gowtham.s
Comments (0)

Please sign in to leave comments

Physical and intellectual resources needed (besides financial resources)

Support from any AI experts while building the model.

March 30, 2020 at 10:21 AM
Created by amudha
Edited by gowtham.s
Comments (0)

Please sign in to leave comments

Next steps? Pilots?

The steps I propose to follow to create the model are:

1. There are all 3 kinds of data present that is Structured, Semi-Structured and Unstructured data in our database. We first categorize them by applying clustering technique so as to find their semantic tags. Thus clustering them into groups is a straight forward method to group them according to some similarities.

(or)

1. We can use NLP (Natural Language Processing) to find out relation among documents, their entities, the main keywords used etc. and group them accordingly.

2. The above technique leads us to somehow group the vast amount of data and thus we can use AI techniques to build a Knowledge Graph that links Structured, Semi-Structured and Unstructured data.

3. Thus obtaining Knowledge Graph is the most efficient way to do future predictions on those data using AI techniques (Machine or Deep Learning).

March 30, 2020 at 10:21 AM
Created by amudha
Edited by gowtham.s
Comments (0)

Please sign in to leave comments

Cost structure? Financial Sustainability? Revenue streams?

Since all the documents are digitalised and is in the UN Library there aren't any costs for procuring data.

Creation of Knowledge Graphs can take place within local server because using cloud platforms would be costly but using local server would be very much cost efficient but the amount of time it consumes depends upon the server itself.

March 30, 2020 at 10:21 AM
Created by amudha
Edited by gowtham.s
Comments (0)

Please sign in to leave comments

How might this go wrong? How might the problem evolve? What are the legal, cultural and other impediments?

The problem might evolve when the knowledge graph is not created in the way we needed, which can be mitigated by few more iterations and enhancements in program.

Another problem would arise when users can't access the data, which means we have to regulate the API.

The entire challenge is mostly code(software) based hence any problems can be rectified.

March 30, 2020 at 10:21 AM
Created by amudha
Edited by gowtham.s
Comments (0)

Please sign in to leave comments

How will i promote adoption?

The entire reason of creating an API and a search bar for Knowledge graph is to encourage user to use and explore the archive data.

We can launch a free Data Science campaign in any Data Science hackathon websites (Kaggle, Hackerrank, Analytics Vidhya, VLearn) that can use the API we built to do any kind of data exploration or use of any ML/DL Algorithms of their choice or we can give a specific topic like pollution, diseases, etc.

March 30, 2020 at 10:21 AM
Created by amudha
Edited by gowtham.s
Comments (0)

Please sign in to leave comments