TagHub.Dev
As promised in day 0, I’ll share more about this TagHub project until it becomes too mysterious (I want it to be friendly!)
TagHub is a textual analytics platform, just go to these products, see, and you know exactly what we’re building:
You input texts (unstructured), you’re getting out some brief report of the corpus (e.g is it positive or negative, what words are mentioned the most?), automagically.
So what’s the point of making the same product: We’ll make it free. And open-source.
MonkeyLearn costs upfront 300$ for 3 users for a month, for highly functional, easy-to-use and beautiful analytics app.. I do admit that they have good color science for the charts (at least it pops to my eyes lmao). Their targets are mainly businesses so the price is self-adjustable for a valuable tool. However, that’s a bit expensive for, say, a student can afford it for hobby projects.
New member to the team: Frontend
The more the merrier… Frontend engineering has been for the most part the bottleneck of this project: We don’t inherently have a key member for the frontend side, the only guy who coded frontend so far is me, who was a ML engineer. Despite of having some axonal confessions of love for front-end designing, my frontend skills are no where near production-graded. It was til the point I was so desperate that I suddenly receive this notice from a frontend guru, who is my friend in a English learning club back in the day at university.
This guy is named Bach, who is intrinsically a funny guy. He, however, does not frequently talk much (aka exchange pleasantries) outside of his scope of work. He has a highly critical mindset, with some drawing skills, definitely a perfect fit for the frontend role. I’m just so glad to have him on the team.
We’ve decided to meet each other at a coffee just after 5 minutes exchanging info about this project with messaging app of the evil corporate, messenger. Mark Zuckerberg was right this time to leverage his AI Machine Learning sh** to inform my friend that I have a project running and need a frontend guy asap. Therefore I would highly suggest you guys star the recommendation system repo from Facebook.
I believe we had a very productive talk. This guy, you tell him the idea and he tell you exactly how we’re gonna implement it.
Design Language Picking
After the talk, Bach helped collect some of the design from various websites/products to choose which design language we’ll opt for later. It was no easy task, to me every frontend pieces he gave just look awesome..
The list is as follow:
- Codecademy
- Observable
- Brilliant
- Figma
- HuggingFace
- Miro
- RayCast
These are websites with good designs. If you go to this website, you’ll see that heir picked design language was not for just the sake of having a working webpage, but to convey a message & artistry of the product.
Not an expert but through quick look, I have hearted eyes for Miro and HuggingFace for being the two simple yet elegant looks, no glitz and glamour, just plain functional. On the other, RayCast, for example, is gentler and fancier, but it’s too much of a design.
… Til next time