Blogs & Past Meetups

Starting a Career into Data Science A Roadmap for Skills in 2022

Updated: Aug 9




Written by the Founder Karachi AI - Mehsam Raza Hemani
I was reached out on Soul Brothers Facebook Group for assistance, so I wrote a comment here is a formatted version of the advice to get started into Data Science

Those who want to start into Data Science should acquire skills in the below sequence:


1. MS Excel

The main tool where most of the Data Analysis is done, learn how to Pivot Data and make analytical summaries along with Basic Charting.

Time Required:

- You will learn this tool by practice in a 1-2 Months

- You will master this tool by practice on your job all your life.

Outcome:

- You will acquire Logic Building & Analytical Thinking in the above Process but at Basic Level


2. Databases & SQL

In short non-technical its MS Excel on Steroids means large capacity, Start learning about them and SQL is the way to retrieve and analyze data from Databases.

Time Required:

- You will learn this tool by practice in a 1-2 Months

- You will master this tool by practice on your job all your life, while you get access to database. (Other way around is to use online datasets and platforms)

Outcome:

- You will acquire Logic Building & Analytical Thinking in the above Process but at Intermediate Level


3. Business Intelligence and Analytics

Start learning Power BI for Data Analysis and Tableau for Story Telling, in the Business world you will be required to Analyse data in these tools and create reporting Dashboard for Daily, Monthly, and Quarterly consumption as per Strategic objectives and Company needs

Time Required:

- You will learn this tool by practice in a 3 Months

- You will master this tool by practice on your job all your life, just download and start using excel files as a source and create Dashboards and Reports for your company. Business Metrics, Domain based Analytics is the key knowledge to derive value in this area. Eg. Metrics for Oil Gas will quite different from Metrics for Investment Company.

Outcome:

- You will acquire Logic Building & Analytical Thinking in the above Process but at Advance Level




4. Programming & Coding Skills :

Now your use cases will be more around creating Custom Solutions and More Free hand logic Building per your Data Analysis or solution Building Cases. Welcome to Python the most widely used and easiest to adopt language for Programming and Scripting Data Analysis and Building Scalable Solutions.

Time Required:

- You will learn this tool by practice in a 3 Months atleast or 6 months.

- You will master this tool by practice on your job or if your job doesn't just pickup a free online cases study or dataset or a use cases and start solving the problem with Python. You will build solution over time, and every time you do this again and again your solution and efficiency and proficiency is improved. You will learn new things along.

Outcome:

- You will acquire Logic Building & Solution Building in the above Process and will become invincible to code anything and make it a reality. It is a really powerful skill I rejoice as a Non-CS Grad.



5. Modern AI & Machine Learning :

Not all problems will be solved by Coding a Logic and you will find a lot problems in this world not scalable to be solved by a logic. Welcome to modern way of writing a Software, using Machine Learning to Learn the Logic from Data and Produce expected Output. We can use this ability to Predict Output Against an input Such as - Forecast (Regress), Differentiate (Classify), Recommend (Cluster).

You will start learning Machine Learning Techniques, this will leverage your skills in Python as well as BI tools as now many tools are coming AI infused too where you can just use them drag and drop and no use of code as far as you understand the concepts. (This is an entire area where people even specialized to become Machine Learning Analysts / Engineers, such as NLP Engineers, Computer Vision Specialist, Algorithm Experts)

Time Required:

- You will learn this tool by practice in a 3 Months atleast or 6 months.

- You will master this tool by practice if you get a Machine Learning Role only, Currently people are hiring for Data Science but giving the same work of Business Intelligence to those hired.. So ML based work hona zaroori ha. Sometimes it also happens you start with BI and then you're giving more advance projects involving ML.

Outcome:

- You will not be able to master it before you practice this in production since its a new area and there is a lot happening upgrading so forget about mastery but be more focused over delivering the results. Not Fancy solutions but cost effective production oriented small solution is a lot better. A simple use case that I have enabled my company to hire better by using ML on the data of past employees who were performant and use their traits to Train a model that approves a candidate with the highest potential and traits of a winning candidate. Its a billion dollar solution in production.



6. Data Engineering and Cloud Computing Skills

Once your done learning the techniques and skills, you will realise that some technologies, techniques and architectures don't scale. Data Science must come outside your Laptop and be delivered to the world at Global scale. Now you will start learning large scale techniques to store and manage data, write efficient pieces of code and assemble large architectures that are rock solid to deliver at any scale and for the scale you will find cloud computing on the top to deliver all the results and lightning speeds we see around our daily apps such as Daraz, Careem, Uber, Airlift. They all run on AWS GCP Azure, so you got to learn the Cloud stack to scale your code from Lab to Production.

Time Required:

- You will learn this tool by practice in a 6 Months atleast or 9 months.

- You will master this tool by practice if you get a Data Engineering or Technical Cloud Associate Role only. However at your scale you can adopt best coding practices, use better technologies and it can become the derive of your own and you can learn by implementing along the way. I have learned these engineering techniques the same way and so can you. Even if you're not much hands-on you must know the eco-system to understand enough to manage a team underneath.

Outcome:

- You will be able to write a better code, execute things better at scale and as you master your skills and become proficient, in 5-6 Years if you're consistent then I know people who have been even hire internationally and even by renowned FAANG group too. These company required talented and efficient Problem Solvers.



In a Nutshell:

This is not only Data Science, Data Science is all the Eco System Stated above 1 - 6) and hence the one who knows it all is known as a Data Scientist.. But Again.. There are roles Jr. Data Scientist, Data Scientist, Sr. Data Scientist, Manager Data Science, Director Data Science. The roles vary as your acquire Experience, Skills and Domain Knowledge that can be applied to the real world to make useful decisions and improved outcomes.


7. These skills or ingredients will be required all along:

- Being Passionate (Not because your friend is earning a 100K in this field, it should be because you love doing this and solving problem with Data Analytics, If you're an Artist then why would you become a Data Scientist? A Data Scientist makes a 100K for example a Passionate and Innovative artist makes 100M)


- Being Curious all the way and being Persistent, Learn the art of "Googling" your query when you're stuck and start loving the problem. Data Analytics and IT Field is all about problem solving so if you're hesitant this is not your type.

  • You will find most of the Problems to be solved:

  • Business or Domain Knowledge : Write your problem and try to search them with "Gartner", "KPMG", "EY", "PWC", "Accenture", "Deloitte", "McKinsey".. For Example How to Measure Supply Chain Mckinsey Report" for getting Knowledge around the subject it will revert with an article or report that you can read.


  • Technical Logics and Programming Questions : Write your Problem, "How to remove last item from the list in Python" and you will find a lot of websites pop-up but one of the most prominent all developers use is "Stack-OverFlow", Read through the question, read the answer with a Tick marked and then analyze the code for your learning to develop the idea.



"I hope my time dedication and knowledge sharing is useful for you and you must have learned from this Article"
About the Author:

Mesum is the Founder Karachi AI , by qualification he is a Oxford Brookes Grad and an ACCA who turned his back on conventional practice and diverted his career along the lines of Analytics, & Data mining. He has specialized in AI & Machine Learning and holds numerous Community Leadership positions in Data Analytics and AI Space. He is a passionate Data Professional and have taught more than 1000+ Students in-person around Data Skills under the Academy banner of karachidotai.com


Thankyou!

181 views

Recent Posts

See All