As one of BfE’s newest teams, Data Analytics has yet to finalise its goals. From the outset, we thought there would be two main goals; using clever analysis to help start-ups and streamlining the internal operations of BfE via insights from our internal data. Given that most start-ups, working together with BfE, do not have much data, we prioritised improving our internal operation.
Having first conducted an analysis of BfE’s past work, we realised one of the main bottlenecks was data collection. BfE produces a lot of data during the start-up programme, including consultancy briefs, call summaries, questionnaires, and presentations. However, the storage of this information in Word documents impedes the extraction of information, linking to other data and final analysis.
Our BfE Tracker project changed this, developing a platform where all of BfE’s data from engagements can be stored in a more structured form. The Tracker not only simplifies the search for information and relations between different documents, but more importantly facilitates the export of data for analysis. We tried different tools, gathered feedback from our different teams within BfE, and developed the prototype.
--By Tomas: BfE Tracker Project
After being introduced to Python 2 years ago, I have been primarily focusing on machine learning. However, our use of datasets under a learning environment for the majority of codings mean they are not ‘real’, so, the ability to work on real data, rather than machine learning, has been fascinating. Flask has particularly intrigued me with its capability to build web applications and I hope to learn and progress much more with the team.
Building web applications in BfE demands a different mindset compared to a normal ‘school’ environment where you focus on solving your own problems. Here at BfE, we think from the user’s perspective to ensure our web application is as simple and user-friendly as possible, as well as having the relevant features. Having acquired a lot of experience so far, I will only continue to improve my technical skills in this area.
--By Singapore Team: Investor Database Project
Capitalising on heritage data and facilitating the selection process for investors, our team built the Investor-Selector system.
When the Finance and Consulting team searches for available funding sources for start-ups, a major task is to sift through large numbers of possible investors. The process is rewarding, yet it would be convenient if we could easily reuse the efforts of previous members. That is why we conceived the idea of building an automated system, which, based on previous members’ work, does a round of pre-filtering and shortlists the potential investors for our consultants to focus on.
Though challenging, the journey of constructing this system has been very fruitful. The task involves multiple exciting aspects of data science from data crawling to text data analysis. So, we are more than glad to add a taste of intelligence to BfE while gaining valuable experience and knowledge in our own professional fields.
Our vision is to use data technology to improve BfE’s efficiency. By focusing on digging values from the BfE heritage data, we are currently building accessing tools, like Investor-Selector, whilst quantifying and visualising BfE’s growing impact.
--By Xiaodong: Investor Selector Project