interesting insight from boom

I listened intently to Blake School on the pod. I figured a good takeway was his quote, where “the amount of effort to run Boom is about the same as his mobile app company”.

Having a problem set that on the surface looks much harder (such as supersonic travel at Mach 2.2) requires just as much energy as solving problems in a mobile app company.

thinking about a new space

I’m thinking about expanding into a new area. the initial evaluation seemed positive. the pros outweigh the cons.

My one main concern is that it takes away a certain amount of energy to be doing this.

The upside is huge, because it brings in a new segment, and improves our overall visibility.

branch out or stay the course

I find myself being pulled into different directions. These do not have immediate impact, but I can see a good outcome if enough effort is channeled into it.

Often times i get into a ‘false dichotomy’ thinking. If I do X, I won’t be able to do Y *that well, because my energies are spread out.

On one hand, it forces one to focus, but it increases the missed opportunities.

I also think it’s possible to scale oneself by hiring more people to help out. While maintaining the core operations and pumping in more R&D. All this comes with stable management and resources.

I’ll put a bookmark on this, and keep thinking about it. I believe there is no one-size fits all solution, more of a case by case basis.

starting from null point

Let’s say one starts out with nothing. No product, no funds, no nothing.

A client approaches you and says ‘Can you deliver solution X’.

You’re confident you can. How do you work around all the initial limitations?

  1. Build a prototype
  2. Demonstrate to build trust
  3. Raise funds via first sale contract
  4. To bring in more funds, raise from outside investors
  5. When the funds come in, start building the full product given a delivery date

gaining an appreciation for data

I’ve come to appreciate data more as I learn about Machine Learning.

The use cases for modelling real-world data and making predictions are plentiful. I can see why data insights are so valuable for organizations and businesses around the world.

educational space

The education space is a massive industry. From schools to materials to vendors to playhouses to extra-curricular activities, the playing field is gigantic.

I’m interested in identifying the performance metrics that matter for educational institutions.

I’m also deeply interested in identifying the problems that parents, teachers, kids, administrators and vendors face, day-in and day-out.

I think there’s something interesting that can be done, to solve these problems. Perhaps by leveraging tech, or simply by optimising certain processes.

true positives, true negatives, false positives, false negatives

While studying the Pima Indian Diabetes dataset with machine learning algorithms, I re-discovered something I learned earlier, which I didn’t quite understand. Now it’s crystal clear.

False positive = Type 1 error. Cases where the classifier falsely predicted a positive.

False negative = Type 2 error. Cases where classifier falsely predicted a negative.

Screen Shot 2016-07-11 at 9.26.21 PM Screen Shot 2016-07-11 at 9.26.17 PM

I can think of this in a non-medical setting too.

1) Classification problem: assume the scientist is traversing a large dataset of companies (potentially millions of rows of companies formed since the 1990s). Feature names could include

  • the number of years the business has survived,
  • how many founders
  • how soon to profitability
  • legal threats
  • location
  • early stage funding

The response vectors would be ‘still in business’, or ‘out of business’.

A false positive would mean given a certain feature set, a company is actually out of business. But in fact, the model predicts wrongly that it’s still surviving.

A false negative would mean given a certain feature set, a company is actually in business, but the model wrongly predicts the business does not survive.

Having large Type 1 and Type 2 errors such as these could could potentially hamper/cripple investment strategies early stage investors or venture capital funds.

Other examples, where a business objective drives certain metrics to be minimized, or maximized

2) In fraudulent transaction detection (1 is fraud, 0 is NOT fraud), priority is to minimize false negatives. False negative means a transaction isn’t fraudulent, but is detected as one. Hence, website owner might lose a sale! Because the system will flag it and ban transaction from happening.

EDIT: obviously, we should never quickly flag / ban transactions. I believe there should just ask for quick verification from users (such as what credit card companies normally do)

3) In spam filters (1 is spam, 0 is NOT spam), we want to minimize false positives. False positive means an email is marked as spam, even though it isn’t. The email user might miss very important emails!


documentation matters

just finished training one of our teammates on operations. Having a shared google doc with step by step instructions really helped.

Now i’m able to remove myself from the task, because my teammate is able to handle it with confidence.

In the event that this teammate is replaced by another one, she would be able to do just the same.