Mandar Joshi Aren't there any Floral websites already? While it was a fun and entertaining journey to identify plant species, I found that all the relevant information about the species is not aggregated on any single website. I had to refer through multiple websites and books to understand all aspects of species. Although the whole ID process relies extensively on differentiation, there is no website that provides a feature to compare different floral and plant characters with photos. IDing is effectively a drilldown process, refining the filter criteria to reach to the desired species. Hardly any website provides such an intuitive search feature. Botanical jargon is another hurdle to decipher even for a seasoned player. There has to be some easy solution.
As a Finance Business Analyst, I often perform similar search and compare tasks for financial data. e.g., comparing two equity stocks or mutual funds, searching for less risky and more beneficial stock to invest. If we have websites to compare and search different mobile models, why cant we have one to compare and search different plant and flowers?
So, my idea was to build similar website as we have for stocks/mutual funds or mobile comparison, but for plants and flowers - comparing their features and photos.
Hold on - Why not Google Lens? If plant identification is the ultimate goal, why not just click the plant and scan it in Google lens and get a plant name in few seconds? This is my explanation -
There are many species with similar looking flowers (Jasminum - मोगरा) or similar looking leaves (Acacia - बाभूळ). In those cases, flowers, leaf, bark or fruit needs to be looked in combination to identify exact plant name. Moreover, even the same species can show variations due to different soil, water and whether conditions. Plant identification based on just one image (as what Google lens do) is contrary to the science of plant identification perspective.
Google lens works on AI algorithm by scanning billions of labeled images available on vast expanse of Internet. Plant images just happens to be a very small subset of a huge image database. It is a generic image recognition tool, not specifically made for plant identification. Even this small plant database is rather biased towards western plants. Most likely Google lens will predict a similar western species when Indian native species is presented or can predict an altogether wrong species.
In spite of these drawbacks, Google lens still does a good job in predicting most of the garden plants. There are like 30 species of Ficus and more than 100 species of orchids. If you are that nerd, who wants to know exact name of orchid in your garden or exact name of hibiscus in the nursery, you can't rely on Google lens!!
So back to building website -
Software languages & tools used: As technology is becoming more and more commoditized and open sourcing a new norm, different components of website are freely available and there is almost nothing that you need to build from scratch. In such case, ability to search and adapt those components to your needs is a crucial competency. After about 6 months of exploring tech landscape and actually creating a product out of it, I can call myself 'Code Assembler'!
Making website up and running:: I never ever thought that, I could host a website on my own. This was way out of my box, until I attempted it!! Here again key is to navigate thru those infrastructure, platform or software providers, go thru their documentation, try out their free packages and adapt those services to your needs. Have HUGE patience by the way - संयम, धीरज, பொறுமை, సహనం, Geduld, υπομονή, 忍耐.....
It is indeed satisfying that I could crawl first few steps and complete the first mile. This is just the beginning and there is a long way ahead!!
