ShareX

A Knowledge Base for Accountants and Auditors, powered by NLP and AI

Comment

Inspiration

In our 4-man team, two members have pursued internships with auditing firms and government agencies and found that during the auditing process, there was significant time spent on technical research. For example, one of the members, who had done an internship with the Ministry of Finance, had found that he had to search through the web for a document detailing the benchmark of a shared services center for an efficiency analysis. He ended up with a document published by a ministry in a South African District. In addition, he had also encountered a startup building a knowledge base for legal firms using knowledge graphs and powerful natural language models. Synthesizing the former problem with the latter solution, we formulated the idea: a Knowledge Base for accountants to intelligently store, organize and share knowledge among one another.

Thus, we present to you: ShareX.

What it does

Organize, Label and Learn

Organize: Use machine learning to organize and give meaning to your knowledge. Your assets will be utilized to enhance the service you provide.

Label: Automatically labels documents you upload by comparing the words inside the document to our database and using Google's powerful Universal Encoder.

Learn: Our knowledge graph is able to construct a network of words and concepts and maps the relationships between each concept, providing new insights that could be leveraged for value-added work. Knowledge graphs have already been widely used by search engines and databases in big tech companies like Facebook and Google.

How we built it

The front-end was constructed by our full-stack engineer Nigel. It is hosted on Netlify and uses HTML, CSS and React. It also hosts a ChatBot for users to send and refine their search queries and how to navigate the site.

Our back-end is a Flask server which hosts the APIs for our front-end to send its POST and GET Requests to. The Universal Encoder is also hosted on our server and used to extrapolate the best and most suitable keywords for user queries which contain keywords not found in our database.

Our database is a MySQL database hosted on _ Digital Ocean _ and is where we fetch our data from using the back-end before sending the results in the form of JSONs to the front-end for our users.

Challenges We ran into

One of the main challenges we ran into was regarding the feasibility of the whole project. We spent much of our time brainstorming for a realistic idea for a platform that would be would be able to develop, and at the same time reconcile that with the possibility of a lack of demand for our platform product. Our mentors were divided over our idea initially too, with some agreeing that such a platform would be of value to the accounting and audit industry, and others of the opinion that such an idea would not be able to take flight because of the low probability that a knowledge sharing platform would even sustain in an industry like this. However, after refinement and much discussion, we managed to scrap components that we deemed improbable of working out, and instead building on the strengths of our platform to create something with our tech know-how. Incorporating our ideas into the back-end and front-end dashboard was also fairly challenging as we ran into many bugs whilst developing it to a workable prototype. A tight working deadline was also a contributing factor to what made this challenge a challenge, but eventually after numerous hours we are proud to present the next revolutionary platform for the industry, ShareX.

Accomplishments that we are proud of

From this challenge, what we are most proud of was being able to put our skillsets to use, and contribute our strengths individually to create a holistic piece of work. Each of the components of building this platform require different knowledge and skillsets in order to successfully piece together, and seeing how all of this came together for the final product was extremely satisfying, and we can safely say we have put out best foot forward in creating this platform that we sincerely hope will play a role, however significant, in changing the accounting and audit landscape of Singapore. Of course, we did not just give up.

We also had many takeaways from this challenge, and that is something to also be proud of. We entered this challenge with the mindset of growing as developers and also expanding our knowledge of industry rather than just implementing something that we already think we may know. However, after going through this challenge and speaking with our respective mentors, and even consulting other contacts and forcing ourselves to pick up the skills needed to build in a short span of time, we found ourselves enjoying the process. Last of course, we are most proud of the product which we have developed for the accounting and audit industry of Singapore, which hopefully will be able to tackle the problem posed by the sponsors.

Do give the platform a try, and hopefully you will see its value and future potential as we look to scale it up, as have we right from the beginning!

What We Learned

Through this challenge, we have the opportunity to learn many things and these are our 3 key learning points:

Firstly, coming from different majors, specializations and having different skill-sets, we have to learn to be able to complement each other and play to our strengths in order to produce a quality piece of work. Having different work styles further pushed us to learn to accommodate one another.

Secondly, we also have learned how we can creatively tap on new technology and apply it in the solutions we have for the problem we are tackling. Those from accounting background would suggest ideas while those with a background in computing would look at the feasibility of the solution with the implementation of current technology.

Lastly, from the mentors-on-site, we were able to learn about the industry and the challenges that industry professionals face. This rare opportunity, with the gathering of professionals from various key players, provided valuable insights which could not be easily attained apart from interacting with the professionals in person. We were privileged to be able to learn more about the industry's situation and providing ideas which would benefit the accounting scene in Singapore.

What's next for ShareX

1) Construction of the Knowledge Graph using GRAKN.AI and other open source toolkits. 2) Refinement of Keyword search using more accurate Natural Language Processing Techniques 3) Refinement of ChatBot Feature 4) Collecting more data from databases containing relevant and important documents to accounting and liaising with accounting companies to offer our service of organizing their data. To be clear, we won't release our client's data to our people but we will obtain our data from databases on Bloomberg and other subscription services. The costs from these services will be factored into our business model. The client's data will be organized within a separate database and kept within the company while using our services to intelligently store and organize the documents.

You can try out the site at here: https://sharex.netlify.com/

Share this project:

Updates