S
SocialCops Blog
blog.socialcops.com · Company blog · on HN since 2015
SocialCops is a data company in India whose blog covers machine learning in Python, Airflow, Kubernetes and data tooling.
What HN saysHN commenters link it in discussions of serving users in India and of how representative survey samples are; they give little opinion of the blog.
15HN people shared it
2linked it in comments
32stories submitted
685story points
8 years agolast shared
2015people per year2026
Most-shared posts
- 1Would You Survive the Titanic? A Guide to Machine Learning in Python4 people · 12 points · first shared 2016 · discussion on HN
- 2A Python Library to extract tabular data from PDFs2 people · 510 points · first shared 2018 · 99 comments on HN
- 3A Tale of Cloud, Containers and Kubernetes2 people · 5 points · first shared 2018 · 1 comments on HN
- 4How to create a workflow in Apache Airflow to track disease outbreaks in India2 people · 4 points · first shared 2018 · discussion on HN
- 5Airflow, Meta Data Engineering and a Data Platform for India1 people · 58 points · first shared 2018 · 5 comments on HN
- 6Announcing flyio, an R package to interact with data in the cloud1 people · 45 points · first shared 2018 · 6 comments on HN
- 7The Data Science Algorithms That Silently Power Uber Cabs1 people · 6 points · first shared 2015 · discussion on HN
- 8What an Infrastructure Engineer's Presentations Taught Me About Storytelling1 people · 6 points · first shared 2016 · discussion on HN
- 9Using Airbnb’s knowledge repo to scale your organization’s knowledge management1 people · 4 points · first shared 2018 · discussion on HN
- 10Building an Android App Using the MVP Pattern1 people · 4 points · first shared 2016 · discussion on HN
- 11The Internet of Coffee: Our Experience Building a Wifi Coffee Maker1 people · 4 points · first shared 2016 · 1 comments on HN
- 12Think Twice Before Making a Pie Chart1 people · 4 points · first shared 2015 · discussion on HN
- 13What Startups Can Learn from McKinsey About Building a Strong Company1 people · 3 points · first shared 2018 · discussion on HN
- 14What 66 years of global data on coups tells us abt the recent attempt in Turkey1 people · 3 points · first shared 2016 · discussion on HN
- 15Why use the Realm mobile database?1 people · 2 points · first shared 2015 · discussion on HN
What HN people wrote when they linked it
sbetageri · 2017-09-22
As a former android developer in India who asks how to serve the masses, the primary concern comes in functionality followed by design.
I will be speaking from the perspective of India.
Here, the vast majority of the rural users do not have great smartphones. With the recent release of Jio, internet speeds have improved. Still wouldn't call it great, but it is good.
Some of the most important functionality decisions that we make is offline support. We always assume that the network will not be consistent. Google Maps allows certain map locations to be saved offline. If memory serves, this was a feature that was aimed at the developing world but found great uses everywhere. Especially when hiking in locations with poor cellular connectivity.
Our apps are primarily in English. Having the ability to view the same content in the regional language is an added bonus. At the same time, we don't expect everyone who uses the app to be familiar with all the icons that are used.
The size of the application is just as important. Before the launch of Jio, data usage was capped. Back in 2012-ish(when I got my first android phone), I was using a 3G data plan which gave 130MB for 14 days. So we aim to reduce the size of the application.
Coming to design. For first time mobile and internet users, we aim to reduce the number of clicks or taps. This is just good design in my opinion.
A short tutorial goes a long way in providing a good experience.
The following blogs do a much better…
shkkmo · 2019-04-13
> ie. Isnt this just a convenient excuse to deny that a survey is meaningful?
Nope, not at all.
It is true that no sample will ever be perfectly representative of a larger population. However, some samples can clearly be more representative than others and the easiest way to tell is to look at sampling methodology. Sample size has absolutely no effect on removing bias.
Here's some info so you can learn more about sampling methodologies:
https://blog.socialcops.com/academy/resources/6-sampling-techniques-ch…
Now, doing actually representative samples is HARD in many situations, so is knowing how representative your sample is. This is why we can't predict things like who is going to win elections.