Troubled Setting Up Your Own Blog? Then Try Out What I did.

Jarmos — Thu, 02 Apr 2020 10:41:21 +0000

Earlier the other day, I shared my thoughts on why Data Scientists should try their hands on blogging. You can read more on it at As A Data Scientist, What's Your Excuse to Not Blog?. It took me a while but it eventually hit me.

Stating how to do it might benefit more than blatantly stating why they should do it instead. So I decided to share this article as a "hand-held" guide to setting up one's own blog!

I used Hugo which is an extremely fast Static Site Generator & also one of the most popular ones, on par with Gatsby & Jekyll. In fact, Netlify ranks it in no.3 as the most popular SSG out there!

So Why Hugo? Here Are My Reasons:

As a Data Scientist, I want to abstract my web development workflow as much as possible. This way I'm focused on delivering DS/ML stuff than actual Web Dev stuff. If you're comfortable with CLI, Hugo will be your best friend as you'll see further into the post. Trust me on that!
Hugo was built on Go, which gives it an unparalleled production speed. Check out David's Performance Test on freeCodeCamp where he got whooping 8.7 seconds load speed on a 2G connection. I wondered if all I want is to deliver reading resources for my audience, it doesn't matter if I've a login system or a backend system, etc. A simple easy-to-setup, fast & static blog is all I needed.
There're no additional dependencies to download onto my computer! Yes, you read it right. I didn't even have to download Golang to install Hugo. How cool is that? Simply download the required binary, install to $PATH & you're good to go.
The list Hugo themes rival that of the ones available for Gatsby. Coupled with the ease of use, quick development time & modern-looking themes I couldn't find any reason to look elsewhere. Read ahead & I'll definitely suggest some themes for you to check out on your blog.

Now The Guide You've Been Eagerly Waiting For.

Downloading & installing Hugo locally is fairly straightforward. Also it's very well documented in the official Getting Started guide, so definitely check it out. Regardless, here's an overview:

Install Hugo
Create the site locally using the hugo new site <DIRNAME> command.
Choose a Hugo theme & add it as a submodule inside the <DIRNAME>.
Add content using the hugo new <FILENAME.md> command.
Check if your site is fine as per your requirements using the hugo server command.
And finally, build the static pages using the hugo command.

That's it! You're done, well almost! But rest assured the scary part is far behind you now.

Hosting The Site

So create a <REPO_NAME> repository on GitHub & push your local repository up there.
Login to Netlify & click on "New site from Git" which should take you to the following page:

Click on the GitHub button & select the appropriate repo as the image below.

Under Basic Build Settings type the Hugo command & Publish under the "Build Command" & "Publish Directory" respectively.

Deploy site! Change the URL to something that's more human-readable & that's it.

You're done! Your personal blog was developed & deployed in less than an hour. If it's that easy I really wonder, As A Data Scientist,s What's Your Excuse To Not Blog?

Cover Photo by Sydney Rae on Unsplash

As a Data Scientist, what's your excuse to not blog?

Jarmos — Fri, 13 Mar 2020 07:09:14 +0000

Part of working in a Data Science role is about "storytelling". Yes, you read it right & I'm not joking either.

Working in a Data Science team involves a ton of communication while keeping the context simple, straight-forward & easy-to-understand for the non-technical higher-ups in a corporate setting. In fact, contrary to popular belief, an individual working in a Data Science role would rarely be building state-of-the-art predictive models to identify if an approaching celestial object in the sky is a meteorite or a black hole, pun intended.

Sorry to break to you, if you were thinking otherwise.

Suffice to say, besides coding, a Data Science role demands more than what's usually comprehended. Like, Domain Knowledge, among a few other skills, is often overlooked & considered as an underrated skill set. While quite opposite to the fact, Domain Knowledge can prove to be uplifting for the overall capabilities of the Data Science team.

I've shared an article towards the ends of this post, about Domain Knowledge and its use case. Check it out, I'm sure you might learn a thing or two ;)

Besides, while working with real-life datasets, it isn't uncommon to come across unclean & unprocessed data adding up the additional skillset requirements! For example, "handling missing values, properly" is a task, requiring careful assumptions and meticulous investigations. Failure to do so can allow bias to creep in & might as well make the model useless or not generalize well to real-world scenarios.

Well, unfortunate for us, the thing is, observed values might be missing for a variety of reasons, unknown to the analyst. Hence, it is critical to figure out the exact reasoning behind the missing values or at least if possible make certain plausible assumptions about them.

But, the topic of Missing Values is unrelated to this write-up, So I wrote a fairly detailed piece of article here - Dropping Missing Values? You Probably Shouldn't. I highly suggest that you take a look at it.

Regardless, my point is, if you're working in a Data Science role or if you plan on working someday, then prepare yourself because communicating your findings is equally crucial, just as the rest of your job. So, my advice? Start working on the next Data Science project & blog about your findings if you've not started it already.

Interested in reading more of my writings? Then you might like what I've written elsewhere too!

A short abridged version of this topic is available at - Data Scientists ARE Actually Storytellers
Want to read how your extensive years of Domain Knowledge might turn out to be useful to a Data Science team? Then check out - Domain Knowledge - Perhaps the Second Most Important Skill in Data Science

Besides, would you like to reach out to me? Then feel free to drop a tweet on Twitter or subscribe to my mailing list.

Cover Photo by Corinne Kutz on Unsplash

DEV Community: Jarmos

Troubled Setting Up Your Own Blog? Then Try Out What I did.

So Why Hugo? Here Are My Reasons:

Now The Guide You've Been Eagerly Waiting For.

Hosting The Site

As a Data Scientist, what's your excuse to not blog?