Harsh Singhal - Harsh Singhal

What is an easy and sustainable approach to transition to Data Science?

If you are in undergrad and are contemplating a career in Data Science and Analytics, pick up tools and technologies that enjoy industry wide usage such as Tableau, PowerBI, R, Python and others that are commonly used. Learn from online courses and free YouTube

If you had the chance to go back in time, what would you do differently about your data science journey?

Going back in time is a great way to understand what is it that one needs to do now to change course for the better. I love this trick that makes for great conversations and self-examination. One thing I will change if I can

perspective

How to evaluate a Master's degree in Data Science?

If you keep your notebook close, and your pencil even closer, and make notes and connect the dots, and you do this even for a few weeks (30-45 mins daily), at the end of a month you would have gained a whole lot of perspective and understanding.

perspective

How to know which domain of data science is suitable for you?

All of a sudden, a company will go from collecting very little data to now collecting data from their website (views, clicks), employee data, manufacturing (from machines, processes, employee activity) and any other aspect of their business that needs to be improved.

bengaluru

Consistency is the key to building long-lasting value

Shivannas is a small store in Bangalore on JC road. I used to go there as a kid almost three decades ago with my parents. This was my dad's hangout when he went to college near by. My grandfather used to stop there to

machine learning

What is the "quick start" approach to learning Machine Learning?

Take up one topic from your notes and go research it by googling it, finding Wikipedia articles, blogs, more examples of its usage, read related discussions on StackOverflow and gain as much understanding of the concept as you can.

Q&A

What can be done to enhance my career in the field of Data Science?

Everyone learns differently. But to know what works for you, choose a system that you found successful in school or college. Often we used to sit with a resource (usually a book) and read and take notes. Do the same now. The process is the same but the resources have changed.

Q&A

I am trained in statistics but worry what I can contribute? Should I focus on technical skills more? Please help.

The discussion shared below was part of many Q&A sessions Harsh Singhal conducted with Data teams at various companies and colleges.Analytics is statistics applied with code (code = SQL/Python/R/Java/etc.) E.g., when you identify the average or variance

Q&A

Motivations for staying in a role for longer than a few years?

The discussion shared below was part of many Q&A sessions Harsh Singhal conducted with Data teams at various companies and colleges.People stay and leave for a variety of reasons. As long as you are convinced to stay or go, then your

Q&A

Do I need to have basic knowledge in every tool out there for data science?

If you can create a list of every tool existing in data science, that list itself will be fascinating to share with others. After you have done the research to create such a list, add a few lines to summarise what each tool does

Q&A

How can a business analyst transition to ML? Especially when analysts get little or no exposure to ML in their job?

Participate in Kaggle.com and learn AI/ML on https://www.fast.ai/ (free). Do excellent Analytics work by day for your company and spend your evenings pursuing your AI/ML dream. A resource like www.fast.ai has made it very easy for

Q&A

How to build a data-driven mindset?

A mindset is a result of what you repeatedly do. A data-driven mindset means that you bring data to decision making, repeatedly. Bringing data to decisions implies a lot of things. It could mean developing engaging dashboards. Developing datasets to make it easy for

Q&A

How useful is a master's degree in data science? Is it really worth it? Or can I learn the same things on the job?

First, identify what it is you want to learn. Can you come up with a list of topics and projects that you wish to work on? Then, you will have to research various degree programs and their curricula for overlap. Additionally, you should then

Q&A

How to plan and execute the company's ML project for better business impact?

An ML project, like any other project requires the project team to get together in the early stages to resolve lots of open questions to better scope the project. What is the problem being addressed and why is it seen as a problem that

Q&A

How to choose a path (Data Engineering, BI, Data Science, Product)?

Careers will span longer periods (decades and more) increasingly, so people have to be ready to make changes throughout their careers. Changes that are made in the pursuit of media-driven hype can be difficult to sustain as new hype cycles will quickly replace the

Q&A

How to clean up past projects ?Or call it "legacy" and just move on?

The first step is a detailed assessment. Is there a document that takes up different aspects of the previous project to identify pros and cons? Was there a requirements document that was used as a guide to developing the project that is now being

Q&A

What are the essential skills required for each role in the Analytics domain? And how to pursue it?

If you have a job, use the industry you are in as a starting point to explore your aptitude and interest. You might like to write lots of SQL queries, extract insights, and make interesting presentations with spreadsheets and slides. This is as useful

Q&A

How can I have a successful career in e-commerce analytics?

What does success mean to you? This is a broad question to an already broad question :-) Try to identify what interests you about e-commerce and aspects of e-commerce that Analytics can help improve? So many aspects are involved in making a sale online

experience

Gain relevant experience by creating your own project

Common Crawl (CC) is a very popular dataset and can satisfy a great many text analysis and NLP tasks. If you have an AWS account you can access the CC index files in Parquet format. Follow the instructions shared in this article to get

ireadrx

Exciting NLP and Computer Vision research for Patent Chemistry discovery and analysis

In 2021 I and a bunch of very smart and driven ML/AI engineers are working on developing NLP and CV algorithms to discover and extract interesting chemistry from patents.

commoncrawl

Fast HTML parsing in Python

I love BeautifulSoup library in Python. It is one of those libraries that just work and make your life easy. I recently came across a fast parser for HTML in Python, https://github.com/rushter/selectolax and waiting to give it a spin. The

Analyzing Bitcoin Transactions on Google Big Query

I've written some queries on analyzing BTC transactions data here. This is a broader project of mine to analyze many other publicly available datasets with Google's Big Query. Go take a look.

string similarity

Postgres text similarity with commoncrawl domains

Commoncrawl is a public repository of web crawl data made available for analysis. In this post I want to extract the list of domains crawled, stick them into a Postgres

rstats

Multiple Plots using Ggplot2

I am a big fan of the tidyverse set of libraries, especially ggplot2 While there is a raging debate on the use of base-r vs tidyverse to teach R to beginners, I will choose tidyverse for the convenience and for the connected ecosystem of

sql

Just Learn SQL!

SQL is everywhere. From databases that we are all too familiar with such as Oracle, MySQL, Postgres to Big Data offerings such as Hive, SparkSQL, Presto, Redshift, Athena and BigQuery to name a few.