Get the "Applied Data Science Edge"!

The ViralML School

Fundamental Market Analysis with Python - Find Your Own Answers On What Is Going on in the Financial Markets

Web Work

Python Web Work - Prototyping Guide for Maker

Use HTML5 Templates, Serve Dynamic Content, Build Machine Learning Web Apps, Grow Audiences & Conquer the World!

Hot off the Press!

The Little Book of Fundamental Market Indicators

My New Book: "The Little Book of Fundamental Analysis: Hands-On Market Analysis with Python" is Out!

If you liked it, please share it:

In the Beginning… Documenting the Birth of

In the Beginning… Documenting the Birth of

I have built many small web apps (see for a sample). I have also built large content sites such as, which gets hundreds of visits a day. But is my most formal and calculated effort to date. My current batch of web properties, put together barely make a few hundred dollars on a good month. I am hoping to change that. 

I am tiptoeing carefully and playing this by the book as I want to give it every chance possible to make it out there in the real world. Here are some of my caution points:

  • Does it solve a problem? A big fat yes! Unlike my previous efforts, this project streamlines my Twitter feed work and saves me countless hours every week. This is encouraging as it solves a real problem of mine and may solve it for others. 
  •  I won’t fall prey to early free releases where it is too difficult to backtrack and monetize. 
  • I won’t over-engineer or automate the heck out this until I understand if and how customers react to the tool.
  • I’m really, really, really going to try to do some marketing…

And following Vincent Woo’s (from CoderPad) philosophy on decision theory constraints, in order to embark on a new project, it should do at least two of the following things: make you money, make customers’ lives better, or be easy to do. I think fulfills all three. 

I have already created a series of generalized walk-throughs on the process of getting a smart and ML-based landing page based using the lessons learned here — you can find them on this YouTube playlist. But I also want to document the actual progress around this tool and its launch — hopefully, others will follow along, be entertained by my trials and tribulations, and even learn a thing or two. Let’s go.

What is ? 

The official tagline:

“A good Twitter feed enriches the lives of its readers, each and every day. If that’s not your case, let help you find the perfect content for your community. This AI-driven tool will not only search far and wide for the right stories but will also format each Tweet in the most irresistible way by telling the story in a single sentence. This is both attention-grabbing and appreciated by those in a hurry — win-win!”

An example of a lede from

A Technical Adventure

The moment I decided to take my Twitter feed more seriously, I immediately started automating the simple stuff. Things like keeping track of whom I was following or following back those following me. I got a lot of the simple repetitive tasks out of the way and kept ratcheting up the automation. It wasn’t until a few months down the road that I realized that the pipeline I had built started doing some incredible stuff.

Besides automating growth and tracking analytics, the tool matches articles, news, posts that are compatible with one's community. As I don’t have any customers yet, it mines stories for my community — those interested in bootstrapping projects, ai, privacy, automation, and singularity, and content around those themes abound! But it also formats the content into a tweet by searching the text and finding the lede sentence. Such as a quote or a pivotal sentence to display prominently. 

This entailed reading the content, understanding it enough to pull that most salient quote. IMHO, this is the most important aspect of this tool and it needs to be as eye-catching. Every NLP trick in the book has been used here — from NLTK, TDFIDF, Word2Vec, GloVe, Bert, etc. It has gotten a lot better than when I just started but there is still room for improvement and right now there is still some human ‘vetting’ before releasing but hopefully, it will be worthy enough one day to be fully automatic, riveting and irresistible.


Now for the hard part as marketing doesn’t come naturally to me. Believe it or not, one of my first jobs as a fresh college graduate was working the phones at a survey research company questioning people. I absolutely hated that job! Most of the few calls, those where somebody actually answered, ended in insults or hangups. 

I am pushing myself to write posts specifically on and not hide behind a loosely related but safely generalized topic. I need to get the word out about the tool and not my management or programming skills. Though it feels salesy, I will soldier on.

I will also set up a Twitter account dedicated to (@keepthetalk). There, I will showcase the automated tweet feed on a realtime basis.

Go Ahead, Try It Out

The landing page of houses the latest stories that are clickable to funnel into your own feeds. Go ahead, click on one of interest to open it up in a Twitter edit box for editing and submittal — it's free and fun.

Click on any story, edit it if you want and submit it to your feed

What Comes Next

There’s still a lot of work to be done and plenty more posts to write. If this project doesn’t take off, I want to have the wherewithal to recognize it early on and pivot quickly. In the meantime, I will be stretching my marketing muscle to get the word out. It now feels that the parking lot is far behind me and that I am actually climbing the mountain — is that base camp #1 that I am seeing over there?


Thanks for reading!

If you are looking for branding help or wantå to apply similar techniques at scale, don’t hesitate to contact me at

Follow my latest projects at,, and

Manuel Amunategui

Author of “Grow Your Web Brand, Visibility & Traffic Organically: 5 Years of amunategui.github.Io and the Lessons I Learned from Growing My Online Community from the Ground Up”