Weekly Updates -7
Paul Graham suggests avoiding matters of money and disputes from taking up our mental bandwidth. This week had a bit of both.
B2B customers at Screener
We haven't served B2B customers on Screener because of two primary issues.
1. Data rights
2. Complexity
The raw data for personal use is costly. Data for commercial or website use is 20x costlier. And it still comes with certain restrictions. We wanted to have a clarity on these restrictions before we can serve B2B customers.
We also want to keep our team and code-size small. Adding people slows down software innovation and we don't want to get there.
Yet, B2B customers provide a more stable revenue. They also provide a distribution which is otherwise served by our competitors. We started exploring this B2B landscape this week.
1. Data rights
2. Complexity
The raw data for personal use is costly. Data for commercial or website use is 20x costlier. And it still comes with certain restrictions. We wanted to have a clarity on these restrictions before we can serve B2B customers.
We also want to keep our team and code-size small. Adding people slows down software innovation and we don't want to get there.
Yet, B2B customers provide a more stable revenue. They also provide a distribution which is otherwise served by our competitors. We started exploring this B2B landscape this week.
"People" module in Screener
We have a lot of data with names. Names of people in shareholding patterns, bulk deals, block deals and SAST trades. But the data doesn't have any unique id for each name. We wanted to assign person IDs to this data and club common names into a single person. This can cross-link the data from different sources.
We made some progress on this problem this week. We developed a proof-of-concept and linked the names in the shareholding table. The shareholders table on the company pages uses this new algorithm now.
We made some progress on this problem this week. We developed a proof-of-concept and linked the names in the shareholding table. The shareholders table on the company pages uses this new algorithm now.
Tools and Coffee
Sumy - A Python library for text summarisation
I came across this Python library through Marlene's tweet. It looks pretty slick. I loved reading their documentation for various summarisation algorithms.
KC Roasters - "our coffees are shipped within 24 hours of roasting".
Thanks to Anil for mentioning about them in his newsletter. The good thing is that they also ship small quantities (100gms).
I came across this Python library through Marlene's tweet. It looks pretty slick. I loved reading their documentation for various summarisation algorithms.
KC Roasters - "our coffees are shipped within 24 hours of roasting".
Thanks to Anil for mentioning about them in his newsletter. The good thing is that they also ship small quantities (100gms).