Data Science

Extracting insight from data using statistics, code, and domain knowledge.

CurrentintermediateGuide only -- no course yet

Overview

Data science combines statistics, programming, and domain expertise to extract insight and build predictive models from data. In practice it's an iterative loop: gather and clean data, explore it, model or summarize it, then communicate findings -- most of the actual time going to the first two steps.

What it is
The practice of extracting insight and building models from data, combining statistics and programming.
Why it's used
Organizations increasingly make decisions from data rather than intuition alone -- data science is the skill set that turns raw data into that evidence.
Where it fits
Builds on Python and basic statistics; NumPy/Pandas/SciPy are the core tools used day to day.

Core concepts

  • Data cleaning
  • Exploratory data analysis
  • Descriptive statistics
  • Visualization
  • Communicating findings

Example

Most real data science work looks like this: load data, group and aggregate it, and look at the result -- the sophisticated modeling steps most people picture are a smaller fraction of the actual job.

import pandas as pd
df = pd.read_csv("sales.csv")
print(df.groupby("region")["revenue"].sum())

Common use cases

  • Business intelligence and reporting
  • A/B test analysis
  • Building predictive models

Project ideas

  • Analyze a public dataset (e.g. a CSV of sales or weather data) and summarize three findings

Official references