Interviewing a big pile of data

Overview

Tools: Google Sheets, Google search

Concepts: Spreadsheets, pivot tables, histograms.

Philosophy: Building data reflexes.

Assignment

Find and “interview” a dataset. The Sunlight Foundation list of US open data portals is handy, if out of date. Dataportals.org is annoyingly designed but also helpful. We’ll discuss a bit more about how to find a good dataset in class.

Your analysis should:

Use simple spreadsheet tools to identify maximum and minimum values for some numeric field type
Use pivot tables to do basic data summaries based on a categorical field type
Create some basic charts to help deepen understanding of the data

Your submission should be either a 2-4 minute video or a roughly 600-1,000 word written piece.

If submitting a video, embed the video in your post and submit that.

Whether submitting a video or a written post, you should create a publicly accessible Google spreadsheet with your analysis and link to it in the post body.

Lesson

Data model basics

Talk about rows, columns, and data types.

Data finding basics

I’ll adlib this in class, but will cover:

Google search tricks
Open data portals
Characteristics of a good dataset for analysis

Interviewing a pile of data

How To Interview A Big Pile Of Data by David Eads. (Translations: Simplified Chinese, Traditional Chinese).