Technical Hobbies

Visit: https://dataalltheway.com/

A tutorial website with concepts, methods, and example code on various data science, statistics, and machine learning topics.

Visit: https://www.kaggle.com/rohitfarmer/datasets

Curated datasets for data science education and advancement.

Visit OpenSea: COVID-19 The Deadliest Disease In American History

In September 2021, COVID19 surpassed the death toll of 675,000 of the 1918 Spanish flu pandemic in the USA - becoming the deadliest disease in American history. COVID-19 is caused by a coronavirus called SARS-CoV-2. Spike protein (S) on the surface of the SARS-CoV-2 virus binds to human cells by recognizing the ACE2 protein receptor. This binding facilitates the fusion of the virus to human cells leading to the injection of the viral genetic material into the human cells. Once inside the human cells, viral genetic material utilizes host machinery to produce more copies of the SARS-CoV-2 virus.

This single copy of NFT (24 X 36-inch poster) commemorates the record-breaking deaths in the USA by rendering the spike protein in 675,001 unique combinations of RGB (red, green, and blue) color codes. These 675,001 colors are randomly sampled from all the possible (16,777,216) computer-generated colors. It's a work of several weeks of computation that involved generating hex codes from all the possible combinations of RGB values, rendering individual spike protein structures into unique colors, stitching individual images into rows to cover the width of the poster, and then stacking rows to cover the height of the poster. All the steps were carried out programmatically in the R programming language.

Note: In addition to the PNG file hosted on OpenSea, the owner will also get the poster in high-resolution PDF format. The high-resolution file is password protected and will only be available to the owner via the locked content of this page. A 677-page technical specifications document that describes the technical details and lists all the 675,001 unique RGB (hex) color codes are provided for free - download it from https://github.com/rohitfarmer/covid19-nft.

Note: The pixels in the poster are SARS-CoV-2 protein structures. Stretch the image to the maximum to see the individual renderings. A bigger multicolor rendering is available at https://github.com/rohitfarmer/covid19-nft and also in the technical specifications document.

Visit: https://indiachat.netlify.app/

I have developed https://indiachat.netlify.app/ that lists all the popular public multi-user chatrooms/channels (MUC) related to India on decentralized networks - XMPP, IRC, & [matrix]. The motivation is to promote the adoption and use of decentralized and federated communication platforms. In the current era of mass surveillance by corporate giants, the true essence of the internet is to make it more decentralized and bring the power to control users' data back to users.

I earned my ham radio liscence (technician) on August 4, 2021 with the call sign KC3SNU

• KV3B repeater, 2 meters, 146.955, -0.6 MHz, FM (dual-band HT)
• EchoLink (computer)

1. Radioddity GS-5B with Nagoya NA-24j VHF/UHF antenna and Radioddity RS22 speaker mic.

• EchoLink (Windows 10 and Android)
• Repeater Book (Android)
• HamClock (Android)

Latest: Website under construction.

Bioinfo Bot is a Twitter bot collecting tweets from bioinformaticians/computational biologists worldwide since 2017. Besides tweet collection, the bot periodically searches, follows new users, and carries out trend analysis once per month.

Core infrastructure is running on a Google Cloud Platform instance. Python library Tweepy is the main module used to follow people, collect streaming tweets and fetch tweet-related metadata. The captured tweets are stored in flat files per month. Natural Language Toolkit (NLTK) is used for the analysis. At the moment, tweet analysis consists of producing a list of "Top 20 Words", "Top 20 Tweeters", "Top 20 Hashtags," and "Popular Programming Languages & Frameworks." The processed results per month are presented and stored in the form of a word cloud image and a JSON file, respectively. Word cloud is generated using the wordcloud python module.

At the moment I am self hosting services only for my personal use. However, in the future I plan to offer resources to general public. Some of the services that I maintain on my tiny Raspberry Pi 3 based home server are:

1. NGINX for my experimental website and reverse proxy.
2. Calibre for e-books.
3. Plex for movies, music, and audio books.
4. Lychee (docker) for photography portfolio.