Recently I have spoken to a couple of friends who want to make data-driven decisions for their organizations but don’t know what kind of resources can help achieve their organizational goals.
Before I get down to the nitty-gritty of different data-driven designations, I want to emphasize a couple of things you should consider in the applicant. Business sense, communication skills, and writing skills are more important than you think for these kinds of positions. You need people who can think like a business owner, write crystal clear requirements on what kind of data they want to collect, and communicate the story told by data.
There are three main positions in a data-driven career. Data Analyst, Data Scientist, and Data Engineer. In most cases, Data Engineer becomes Data Architect after 4-5 years of DE experience.
Data Analyst/ Business Analyst
Do you need someone to gather data using different analytics (e.g. Firebase, Google Analytics, Comcast, Kochava, Nielsen etc.)? Need someone to make sure your applications and websites are tracking required data properly? Are you going to use that data to make business decisions by using past results?
You are looking for Data Analyst/ Business Analyst!
DA will tell you, what age group is buying your products. DA will help your sales team determine which demographics to target. Your data analyst can also tell you what is causing recent revenue growth.
DA Required skills/Tools:
Programming: For DA, it depends on the company. In many cases, you don’t need a coder. Experience with vlookup and pivot tables in Excel is good enough, but it is good to have some experience with Python/R/Matlab.
Data Munging and Data Extract, Transform and Load (ETL).
Statistics: Analyzing Primary data is good enough. You have a great candidate if he/she knows how to build a statistical model.
Machine learning: He/she should be (at least) curious about it.
Data Visualization tools: This one is easily teachable but it would be nice to have experience with them (Tableau, Google charts, infogram are popular ones).
For the majority of companies, for data analyst positions, statistical modeling and ML are not required skills.
Data Scientist
Are you looking for someone who can address open-ended questions? Do you need someone to answer questions like:
Is it X or Y? E.g., will 1 month free subscription generate more revenue in the long run, or would giving a 5% discount? - Classification algorithms
Is this something weird? E.g., is this email spam? - Anomaly detection algorithms
How much or how many? E.g., How many users will watch Silicon Valley on season 5 premiere - Regression algorithms
How is the data organized? E.g. Which viewers like the same kind of movie? - Clustering algorithms.
What should I do now? E.g. Should I increase or decrease the number of ads I am serving to the user for a higher retention rate? - Reinforcement learning algorithms
He/She should be good at statistics, building ML models to make predictions and answer questions which most of the people in your company haven't even thought of.
DS Required skills/tools:
These are some of the skills/tools to help a data scientist do his/her job. (Reference:- DATA SCIENCE LONDON )
• Java, R, Python... (bonus: Clojure, Haskell, Scala)
• Hadoop, HDFD & MapReduce... (bonus: Spark, Storm)
• HBase, Pig & Hive...(bonus: Shark, Impala, Cascalog)
• ETL, Webscrapers, Flume, Sqoop... (bonus: Hume)
• SQL RDBMS, DW, OLAP...
• Knime, Weka, RapidMiner… (bonus: SciPy, NumPy, scikit- learn, pandas)
• D3.js, Gephi, ggplot2, Tableu, Flare, Shiny...
• SpSS, Matlab, SAS... (the enterprise man)
• NoSQL, Mondo DB, Couchbase, Cassandra... And Yes! …
• MS-Excel: the most used, most underrated as a tool
Data Engineer/ Data Architect
Engineers make everyone’s life easier. Data Engineers make DA’s and DS’s life easier.
Are you looking for someone to install scalable database systems and set up data disaster recovery systems? Someone to build data pipelines which clean, transform, and aggregate unorganized and messy data into data sources/databases?
Data engineers prepare the essential foundation for data analysts and scientists to retrieve the needed data for their interpretation and experimentation.
Ideally, If you don’t have at least 2 data analysts/scientists, there is no point in looking for Data Engineers. In a team of 2, any of them can act as a data engineer.
DE Required skills/Tools:
Data warehousing solutions
Hadoop-based tech (MapReduce, Hive, Pig)
SQL based tech (PostgreSQL and MySQL)
NoSQL technologies like Cassandra and MongoDB
For all of the above positions, requirements can look very different depends on your organizational needs.
Please let me know if you have any specific questions.
https://www.linkedin.com/pulse/what-kind-data-driven-resources-do-you-need-which-career-i-am-it-/