Data Science Team
A practice to set up
Omid Mogharian
V0.2.1 - 06.02.2017
… it’s a mistake to treat data science teams like any old product group ... To build
teams that create great data products, you have to find people with the skills and
the curiosity to ask the big questions. You have build cross-disciplinary groups with
people who are comfortable creating together…
DJ Patil, U.S. Chief Data Scientist at White House
Office of Science and Technology Policy
Roles
Machine Learning
Engineer
Data
Engineer/Architect
Data Analyst
Math &
Statistics
Interpretation/
Visualisation
Modeling &
ML
Math &
Statistics
Machine
Learning
Developing
Developing
Infrastructure
Design
Operation
Data Scientist
Core Team Skills
Relations
Data Science Team
Operations/
System
Administration
Sales
PO/
Customer
relation
BI
App
Development
To make it real
Method
CRISP-DM
The Data Science Process
Communications with Customer
How to? lambdaBig data and Fast data
Big Data Pipeline
Job repository
Scheduler/
Runner
Incremental
Runner
Message QueueData Pipe Agent
A practice for continuous analyse
Severing
layer
Application Environments
Source
Simulator
Stage
Production
Sample Data
Source
Connector
Big Data
Continuous Analyse Application*
Continuous Analyse Application*
To bring accuracy
* Whole software with
several components
which are explained in
previous slide
References
● https://dzone.com/articles/lambda-architecture-with-apache-spark
● https://en.wikipedia.org/wiki/Lambda_architecture
● https://www.mapr.com/developercentral/lambda-architecture
● http://www.kdnuggets.com/2015/11/different-data-science-roles-industry.html
● http://www.kdnuggets.com/2016/03/data-science-process-rediscovered.html
● http://www.datascienceassn.org/sites/default/files/Building%20Data%20Scienc
e%20Teams.pdf

Data science team, a practice to setup