If there is one thing that all students can relate to, is that university consists of non-stop research and endless hours of checking every academic source available. If you have ever noticed or got the feeling that you are hitting a dead end or just not getting what you are looking for altogether, your are not alone.
Throughout her time in at the University of Cambridge, studying for her PhD in Biochemistry, Dr. Vivian Chan was well accustomed to doing scientific research but noticed that with conducting massive amounts of research, sometimes finding the right information was a problem. So what did she do about it? Now having been graduated from University, Dr. Vivian Chan has created an advanced researching platform for scientific research that is unlike any other platform out there today. Sparrho is comprised of over 45,000+ journals that are updated hourly and find content that is relevant to your research, along with information that you may not have known you were looking for.
I was able to ask Sparrho’s creator how her creation is different from the rest and in today’s tangled web of articles, and other internet finds, this type of researching platform could not come at a better time with endless possibilities.
Verge: Throughout your time at university, what was it that bothered you most about looking for relevant scientific literature?
Dr. Vivian Chan: I was working in protein crystallography, a field that lies at the intersection of biology, chemistry, and physics. This meant that the literature relevant to my PhD research was scattered across dozens of scientific databases and journals that were not presented in databases, so it was very time-consuming to search through all of these sources for my topics of interest every week. Furthermore, search tools, like Google Search, are very good at finding specific things that you know that you’re looking for. What wasn’t available is a tool that can give you what you didn’t know what you were looking for.
That’s why we’ve built, and are continuing to develop Sparrho to tackle the discoverability of scientific literature by firstly giving busy researchers a single, intuitive portal for searching through all publications, regardless of publisher and discipline, and secondly training our AI to better recognise both articles that are highly relevant to our users, and those that have a less obvious connection.
V: What was the inspiration behind Sparrho apart from not finding what you were looking for through other academic search engines or researching in general?
D.VC: As a PhD student, I was very lucky to have Steve, a post-doctoral researcher who understood each of our group member’s research niche. Steve would spend around 15 minutes every morning reading the tables of contents from a large selection of journals and with his experience, be able to pick out serendipitously relevant articles to my work, even when the article titles appeared only tangentially relevant. When I found out more about machine learning, I realised that I could ‘digitise’ Steve to provide a better literature discovery experience for scientists.
V: From my understanding, Sparrho works on collecting journal articles, grants, patents, events and any other pertinent research through an algorithm. How does that precisely work through Sparrho compared to other search engines?
D.VC: Sparrho’s search and recommendation algorithms are enriched by our understanding of how our users interact with the content on the platform. Our users from 80 countries, mostly academic researchers, have helped our AI learn which articles and keywords should be recommended by marking items in their personalised feeds as relevant or irrelevant. We believe there’s still a great deal of progress to be made in how humans can work with machines to provide superior recommendations, and we’ve had a great response to our pinboard feature, which allows users to curate and share their own reading lists, and inform our AI at the same time.
V: How far into working toward your Ph.D. in biochemistry at the University of Cambridge were you when you came up with the idea for a better platform to find relevant and current scientific information?
D.VC: I had the idea for what would become Sparrho’s first prototype at the start of my second year as a PhD student, but it wasn’t until I joined Entrepreneur First and met my Co-Founder and Tech Adviser Niluka Satharasinghe, a serial entrepreneur with a background in machine learning, that I had a concrete idea of how I might go about building Sparrho.
V: After you received your Ph.D., how did you go about making Sparrho a reality?
D.VC: I had a problem to solve, but not yet and idea for a solution, so I did lots and lots of interviews with students, researchers, and science enthusiasts to flesh out the right MVP idea to test. Without any capital, I asked friends to help with different parts of building the MVP. Joining Entrepreneur First gave me the time and space to focus on Sparrho, and is where I met many great advisors and my Co-Founder Niluka Satharasinghe, from whom I also learnt enough basic Python for frontend coding to put the final MVP together.
V: What was one of the biggest challenges to make this happen?
D.VC: Defining an MVP to solve a massive, abstract problem was really difficult. I wanted to improve the dissemination of scientific knowledge, but there are many different ways to start tackling such a problem. Even once the MVP is closer to completion, it took me a long time to refine how I communicated what Sparrho was aiming to do, so that everybody could immediately understand why it was important for my product to exist.
V: Sparrho launched in 2013. How have you seen Sparrho grow and what are your hopes for Sparrho in the next five years?
D.VC: Since 2013, we’ve reached users in 80 countries, with 30,000 unique visitors every month, and collected over 46 million pieces of scientific content in our database, which is around 70% of all academic research publications. Our team has also grown and matured, and we’ve vastly improved our user experience, though of course there is always more to be done! Sparrho is designed to make life easier for our users, especially students and researchers with high workloads and tough deadlines.
In the next five years, we hope to have been able to allow experts to curate, summarise, and opine on cutting-edge research, so that their value “by-products” of their day-to-day literature reading is realised for a wider audience. Our platform will give them an opportunity to showcase their skills whilst helping generate exclusive content which is understandable and relevant to professionals in other industries and even the general public. We’re also looking into having diverse content types such as video and audio to better achieve our vision of democratising science.
V: How does Sparrho cope with the amount of scientific literature and research published daily?
D.VC: We have developed distributed, horizontally scalable systems for importing and retrieving data. On the content import side, we can comfortably pull in up to 1 million items daily, although currently we’re only seeing a publication rate of about 4,000 per day. On the content discovery side, we keep all of our data in distributed, indexed compartments which can be split among machines, and if we start seeing user response times dropping below our targets, we can trivially add more machines and the system will rebalance itself without any manual input.
V: How has your journey as an entrepreneur been and what advice would you give other young individuals on pursuing their ideas?
D.VC: Don’t be afraid to pursue an idea that you’re truly passionate about, but you should always question your assumptions with razor-sharp focus. Start first with the problem you’re proposing to tackle – is it a problem that people really care about solving? However difficult, speak with strangers who could be potential customers before you start building anything. But once you’ve gathered a bit of data, don’t get stuck in analysis paralysis. Just go out there and do it! You won’t get it right the first time, so the earlier you start, the earlier you’ll learn from your mistakes. Remember it’s a journey no matter what the outcome is, so cherish the people you meet and establish relationships with along your startup journey.
Stay up to date with Sparrho on Twitter – @sparrho – and enter the chance to win £100/£150 with Sparrho here (free to enter).