Expertise
Data Science requires the right mix of Statistics, Programming, Machine Learning, Visualization, Engineering, and Getting Things Done.
Statistics
Everything we measure and compute using these measurements has a margin of error. We separate the signal from the noise using Statistics.
Programming
R, Python, SQL, Docker, and Spark cover everything from small to big data, research and production, processing and visualization.
Machine Learning
While not rocket science, Machine Learning, and especially DeepLearning (TensorFlow / PyTorch), can definitely fly a rocket.
Visualization
Visualization tools such as ggplot2, Shiny, plotly, d3js, can create excellent graphs with a touch of interactivity.
Engineering
While Research and Agile Prototyping are fun, Production Deployments must be built using best practices.
Getting Things Done
Motivation might provide the spark, but Perseverance drives the steady pace race.
Consult
The path of a typical Data Science project
Know the problem
Data does not exist in a vacuum, it comes with context and human knowledge. Your Domain Specific Knowledge is paramount in answering questions about the data and guiding the project.
Explore the data
A quick look at the data brings more questions, but also removes unfeasible paths from the analysis. Early actionable insights are always welcomed.
Process
Machine Learning algorithms are applied to create and validate forecasts. Optimization, including parallelization, is employed as needed.
Review and Repeat
We have the forecast / prototype / deployment. Does it make sense? Can we improve its performance? Does it scale?
Your next Project?
About
Mike Badescu, PhD
Data Scientist and Founder
Lives in Dallas, Texas
- Open Source projects
- Data Science Mentor at Springoard
Accomplished Data Scientist with advanced quantitative skills in various facets of economics, finance, marketing, and statistics. More than fifteen years of analytical and research experience, both professionally and in academia.