Data wrangling Westworld
MOVING THE NEEDLE
Gauging Election Reactions
Electoral maps received a lot of scrutiny this election, but the New York Times gauge charts (and their jittery needles) were met with a forceful backlash. The issue at hand? Conveying uncertainty in charts. - Visualising Data
- Related: NYT graphics editor Gregor Aisch shares the thought process behind the gauges.
STAY THE COURSE
Here’s What “24 hours” of Running in Boston Looks Like
Come for the beautiful video of runners’ paths unfurling through Boston, stay for the explanation of how it was made. Plus, there’s an interesting bit at the end: some thoughts on tackling the problem of sharing fun proprietary data without compromising user privacy. - Runkeeper
Against storytelling: We need Data Explorers to make sense of Open Data.
“Making data sources available to un-guided analysis is the foundational step of the whole thing. They offer none or few guidance to the user, unleashing a virtually infinite set of interpretations.” - DataPirata
People are talking about: why the polls got it wrong
Last week we took a break from politics. But we do need to talk about polling because there are lessons to be learned by anyone who communicates data. Here are some particularly compelling post mortems.
- Why are we so surprised? (Probably Overthinking It)
- How Data Failed Us in Calling an Election (New York Times)
- Meet a Polling Analyst Who Got the 2016 Election Totally Wrong (Pacific Standard)
- Yes, the election polls were wrong. Here’s why (The Guardian)
New from Mode
Announcement: Want to shape the future of your organization? Start predicting it. We’re teaming up with Twitch to host a workshop about flash forecasting on December 6 in San Francisco. Get all the deets on our Eventbrite page.
Blog: Obsessed with the HBO original series Westworld? We’re right there with you, pardner. During a recent Mode hack day, a group of us built a website full of Westworld data visualizations. For the backstory, check out this post.
Question of the week
At Mode, we use Redshift for the bulk of our analysis. What databases do you use and why?