How to start a Data Science journey
1. Start with appreciation of data, add a little curiosity to understand the data (what business value are the numbers conveying), develop that into a passion for data, reach a level where you are in love with numbers
2. Start with a simple tool like Excel, learn to play with data, learn to automate simple things and write macros (learn VB scripts as required)
3. When you are comfortable with simple small data sets and are able to glean all that you can from that dataset using Excel, move that to a database and learn to work with SQL
4. Learn to work with larger datasets, integrate datasets using SQL or some ETL tools, learn to get insights out of data using SQL/BI tools, learn to visualize using tools like Power BI or D3JS (learn JavaScript as required)
5. Understand what is a data science project like, understand the life cycle, remember that without business domain knowledge the journey of data science is incomplete
6. Learn tools like R/Python and graduate to predictive analytics. learn to apply statistical formulas in excel first then in R and if required in Python, if you want to understand what is happening behind the scenes go deeper into Mathematics/Statistics
7. Learn about distributed storage and distributed computing and if required learn basics of big data (Go as deep as required, learn java if required and other big data tools as required)
8. Learn tools like Spark and SparkML (They give huge performance improvements over large data sets)
9. Learn to work with cloud platforms like AWS and Azure (The future of Big Data Analytics is on cloud)
10. Start following Kaggle competitions for large datasets and problem statements, and keep challenging yourself by competing against the best (and other places like Kaggle)
11. While you are doing all this work on your communication, interpersonal skills and presentation skills (without these you can’t succeed in this continuously evolving space)
12. Most importantly, keep reading, understand your own learning style, there is no one who is 100% knowledgeable, everyone is at some stage of journey, so all point of views are relevant for a learner
13. Last but not the least, understanding the business problem, talking the business logic, identifying what business needs, picking up a relevant and practical business question, mapping it to relevant technical algorithm, interpreting the results and presenting it back to business in their language and following it up and iterating until the desired business impact is achieved - all these make data science come alive and make it useful to all stakeholders
Data science, also known as data-driven science, the term became a buzzword, and is now often applied to business analytics, however, it was coined in 2008 by employees leading their data and analytics efforts at LinkedIn and Facebook. Let me tell you that this 'data science' existed prior to 2008 also!! I discovered it in 2005 (haven't given a name as such!!!) and started learning/working on 'SAS'. ~ I see this as a profound company that can actually synonymize the word 'data science' in all it's applications~. However I feel that the companies in India have'nt anticipated it then 2005, which took a toll on my career!!! I found this useful : https://www.sas.com/en_us/learn/academy-data-science.html
you have mentioned practical way to become a data scientist from computer expert's perspective .. However , basic knowledge of statistics like types of variables,measurement scales ,mediation -moderation , correlation , regression ,exploratory data , confirmatory data , basic knowledge of matrices and vector math is essential ...
In the business of innovation, it is riskier to be risk averse.
7 年You missed the most important parts of Data Science - the scientific method/hard science (statistics) and business aspects. A data scientist is equal parts developer (as you described), domain expert (business knowledge) and scientist. You focused on only one-third of the requirements. To be a data scientist will take everything you describe PLUS the same effort for the other disciplines.
Data Analytics | Data Fabric | Data Mesh
7 年Simple yet makes so much sense.
This roadmap is perfect for one who needs to step in I had been confused a lot when I started my journey to step into Data Science was not sure . Was thinking of steps 6 to 9 more than the initial which I feel are the most important once. Thanks Vinayak Pai for these insights