while working on data projects I was tired that I couldn't easily share Machine Learning insights that could boost my clients’ business performance without a full-stack IT team.
I was tired of how slow it was to show data analytics to people in a professional way with a progressive-web app technology (PWA from now on). The only alternatives: non-scalable BI tools, Github libraries that are nice as a hobby but not to deliver professional outcomes, or stuff such as power points or surrendering to the bureaucracy of slow & expensive product developments with IT.
I was tired that any small change in a dashboard that someone thought was a good idea - adding a new plot, a new component or changing a chart - meant a drama that needed to go through an endless flow of discussions and Scrum.
I wanted data to be agile, really agile. I wanted it to allow me to experiment, make super fast tries, change in a glimpse, adapt and tune in minutes. But nowadays creating data products is slow. Very slow and painful.
As a data scientist I wanted to be able to work end-to-end without needs or dependencies that would block my work until someone else would have the window to do their part. I wondered why data scientists or data engineers could not be autonomous offering scalable data analytics solutions to their users. I wanted to mix all the business requirements and market knowledge with some transactional data that would allow me to offer machine learning insights to any company. And for that I needed many software engineers to support my work.
This gave birth to Shimoku.
I was tired that any small change in a dashboard(...)meant a drama that needed to go through an endless flow of discussions and Scrum.
Download this post in PDF to have it whenever you want